FACTORS AFFECTING FOREIGN DIRECT …eprints.utar.edu.my/698/1/FN-2012-1001227.pdfFactors Affecting Foreign Direct Investment Decision in Malaysia iv ACKNOWLEDGEMENT Without the assistance,

FACTORS AFFECTING FOREIGN DIRECT

INVESTMENT DECISION IN MALAYSIA

BY

ONG KER XIN

P’NG GEOK THYE

POON DAO CHUN

TAN LAY YOKE

YONG KAH CHUN

A research project submitted in partial fulfillment of the

requirement for the degree of

BACHELOR OF FINANCE (HONS)

UNIVERSITI TUNKU ABDUL RAHMAN

FACULTY OF BUSINESS AND FINANCE

DEPARTMENT OF FINANCE

AUGUST 2012

Factors Affecting Foreign Direct Investment Decision in Malaysia

ii

Copyright @ 2012

ALL RIGHTS RESERVED. No part of this paper may be reproduced, stored in a

retrieval system, or transmitted in any form or by any means, graphic, electronic,

mechanical, photocopying, recording, scanning, or otherwise, without the prior

consent of the authors.


iii

DECLARATION

We hereby declare that:

(1) This undergraduate research project is the end result of our own work and that

due acknowledgement has been given in the references to ALL sources of

information be they printed, electronic, or personal.

(2) No portion of this research project has been submitted in support of any

application for any other degree or qualification of this or any other university,

or other institutes of learning.

(3) Equal contribution has been made by each group member in completing the

research project.

(4) The word count of this research report is 21506 words.

Name of Student: Student ID: Signature:

1. Ong Ker Xin 10ABB01498 ______________

2. P’ng Geok Thye 10ABB01695 ______________

3. Poon Dao Chun 10ABB01696 ______________

4. Tan Lay Yoke 10ABB00364 ______________

5. Yong Kah Chun 10ABB01227 ______________

Date: _______________________


iv

ACKNOWLEDGEMENT

Without the assistance, cooperation and guidance of several parties, this

research project would not be achievable. We would like to take this opportunity to

thank everyone who have helped and guided us in completing this research project.

Firstly, we would like to express our heartfelt gratitude and appreciation to

our supervisor, Ms. Wei Chooi Yi who had guided us throughout the duration of this

study. We appreciate the valuable time, guidance and advices she has given us for the

completion of this research project.

Furthermore, we would also like to thank the lecturers and tutors of UTAR

who have guided us directly and indirectly with new insights and ideas on the path of

completing this study. Besides, we deeply appreciate the moral support,

understanding and endless love in which our family have given unconditionally

throughout the process.

Lastly, the cooperation and support received from all the members who

contributed to this research was vital for the success of this project. The ideas,

suggestions, and perspective given greatly enhanced this research project. Once again,

we are in grateful and in appreciation of all the assistance contributed for our study.


v

TABLE OF CONTENTS

Page

Copyright Page…………………………………………………….………………….ii

Declaration………………………………………………………….………………..iii

Acknowledgement…………………………………………………….……….……..iv

Table of Contents……………………………………………………….…….…….…v

List of Tables………………………………………………………………………....ix

List of Figures………………………………………………………….……………...x

List of Abbreviations……………………………………………………………....…xi

List of Appendices………………………………………………………..………....xiii

Preface………………………………………………………………………………..xv

Abstract………………………………………………………………………….…..xvi

CHAPTER 1 RESEARCH OVERVIEW

1.0 Introduction……………………………………………………….1

1.1 Research Background…………………………………..….…..…..1

1.1.1 Foreign Direct Investment in Malaysia……….…..........4

1.2 Problem Statement…..…………………………….…………........5

1.3 Research Objectives……………..…………………….………....12

1.3.1 General Objective……..……………….………….......12

1.3.2 Specific Objectives………..…………….………….....12

1.4 Research Questions………………………..……….………….…12

1.5 Hypotheses of the Study………………..………….………….…13

1.6 Significance of the Study………………………..….……….…...14

1.7 Chapter Layout…………………………………..….…...….........16

1.7.1 Chapter 1………………………………….…….….....16

1.7.2 Chapter 2………………………………….………......16

1.7.3 Chapter 3………………….…………….………....….17

1.7.4 Chapter 4………………….…………….……...….….17

1.7.5 Chapter 5………………………….…….………….....17

1.8 Conclusion…………………………………….….………….......18


vi

CHAPTER 2 LITERATURE REVIEW

2.0 Introduction………………………………………………............19

2.1 Review of Literature…………………………………….….........19

2.1.1 Foreign Direct Investments……………………............19

2.1.2 Economic Growth…………….……………….…..…...20

2.1.3 Market Size……………………………………..…..….21

2.1.4 Inflation Rate…………………………………..…..…..23

2.1.5 China FDI Inflow…………………………….………..25

2.1.6 Exchange Rate……………………………….……...…27

2.1.7 Trade Openness……………..………….….….…….....28

2.1.8 Quality of Infrastructure………………..……….…..…30

2.2 Review of Relevant Theoretical Models………………..…...…...32

2.3 Proposed Conceptual Framework……………………..….….......33

2.4 Hypotheses Development…………...…………………..….….....35

2.5 Conclusion…………...…………………………………..…...…..37

CHAPTER 3 METHODOLOGY

3.0 Introduction……………………………………………..…….….38

3.1 Research Design……………………………………….…….…...39

3.2 Data Collection Methods…………………………………....……40

3.3 Data Processing………………………………………….….…....41

3.4 Data Analysis…………………………………………….….…...42

3.4.1 Eviews……………………………………….……...…42

3.4.2 Multiple Linear Regressions……………………...…...43

3.4.3 F-test Statistic………………………………….…..…..44

3.4.4 T-test Statistic……………………………………...…..45

3.4.5 Diagnostic Checking…………………………….….....45

3.4.5.1 Model Specification and Normality

Test……………………………………..…..46

3.4.5.2 Multicollinearity………………………...….47

3.4.5.3 Autocorrelation……………………………..48

3.4.5.4 Heteroscedasticity…………….…………….49


vii

3.5 Conclusion……………………………..………………………....49

CHAPTER 4 DATA ANALYSIS

4.0 Introduction……………………………………………….…….…50

4.1 Empirical Result of Multiple Linear Regressions Model……........50

4.1.1 Diagnostic Checking of Multiple Linear Regressions

Model………………………………………………….51

4.2 Problem Solving………………………………………….…..........54

4.2.1 Problem Solving of Model Specification………….......54

4.2.1.1 Diagnostic Checking of Semi-Logarithmic:

Log-Lin Model……….……………….………55

4.2.1.2 Problem Solving of Multicollinearity….……..55

4.3 Empirical result of Model 1……………………………….…........57

4.3.1 Diagnostic Checking of Model 1………………..….....58

4.3.2 Heteroscedasticity Problem Solving of Model 1……...59

4.3.3 Autocorrelation Problem Solving of Model 1………...60

4.4 Empirical Result of Model 5…………………………..….…….…60

4.4.1 Diagnostic Checking of Model 5……………….…..…60

4.4.2 Problem Solving of Model 5…………………….….....61

4.5 Interpretation of Multiple Linear Regressions Results of Model 1

and 5………………………………………………….…….…......62

4.5.1 Interpretation of Model 1 Result……………….……...63

4.5.2 Interpretation of Model 5 Result……………….…..….64

4.6 Conclusion……………………………………………….………..65

CHAPTER 5 DISCUSSION, CONCLUSION AND IMPLICATIONS

5.0 Introduction…………………………………………….….……….67

5.1 Summary of Statistical Analyses………………………….….........67

5.2 Discussions of Major Findings………………………………..…...70

5.2.1 Market Size……………………………………….…..70

5.2.2 Economic Growth……………………………….…....71

5.2.3 Exchange Rate………………………..……….….......71

5.2.4 Quality of Infrastructure………………………….......72


viii

5.2.5 Trade Openness………………………………...……..73

5.2.6 Inflation Rate……………………………..……..…....74

5.2.7 China FDI Inflows………………………………........75

5.3 Implications of the Study………………………………..…….........76

5.4 Limitations of the Study………………………………….………...78

5.5 Recommendations for Future Research……………..…..…….........79

5.6 Conclusion………………………………………………..…….......80

References…………………………………………………………………...…..…...81

Appendices…………………………………………………………………..….........88


ix

LIST OF TABLES

Page

Table 4.1.1: Summary of Diagnostic Checking of Multiple Linear

Regressions…………………………………………………………………………..51

Table 4.2.1.1: Summary of Diagnostic Checking of Semi-logarithmic: Log-Lin

Model….......................................................................................................................55

Table 4.2.2: Summary of Comparisons among 4 Models…………………..…….....56

Table 4.3: Summary of Diagnostic Checking of Model 1…………………..…….....58

Table 4.4: Summary of Diagnostic Checking of Model 5…………………..…….....60

Table 4.5.1: Interpretation of the Significant βs of Model 1……….………….…….63

Table 4.5.2: Interpretation of the Significant βs of Model 5…………….………......65

Table 5.1: Decision for the Hypotheses of the Study………………………..………68


x

LIST OF FIGURES

Page

Figure 1.1: Total FDI inflows in Malaysia (BoP, current US$): 1970 – 2010….....…6

Figure 2.1: Theoretical Model…………………………………………………….....32

Figure 2.2: Researcher’s Model……………………………………….......................33


xi

LIST OF ABBREVIATIONS

ADF Augmented-Dickey Fuller

AIC Akaike Information Criterion

ARCH Autoregressive Conditional Heteroscedasticity

ARDL Autoregressive Distributive Lag

ASEAN Association of Southeast Asian Nations

BoP Balance of Payments

CFDI China FDI Inflows

ECM Error Correction Model

Eviews Electronic Views

FD Financial Development

FDI Foreign Direct Investment

FE Fixed Effects

GDP Gross Domestic Product

GDPG Economic Growth

GNP Gross National Product

GRACH Generalized Autoregressive Conditional Heteroscedasticity

GRO Annual Growth Rate

GUI Graphical User Interface


xii

INF Inflation Rate

INF Infrastructure Development

LM Lagrange Multiplier

MFDI FDI inflows in Malaysia

MNCs Multinational Companies

MNE Multinational Enterprise

OECD Organisation for Economic Co-operation and Development

OLS Ordinary Least Square

OPE Trade Openness

OREER Official Real Exchange Rate

RE Random Effects

RER Real Exchange Rate

RESET Regression Equation Specification Error Test

SAARC South Asian Association for Regional Cooperation

TAX Corporate Tax Rate

TL Quality of Infrastructure

TO Trade Openness

UNC Macroeconomic Uncertainty

VIF Variance Inflation Factor


xiii

LIST OF APPENDICES

Page

Appendix 1: Empirical Result of Multiple Linear Regression Model………..…..…88

Appendix 2: Normality Test of Multiple Linear Regression Model (Jarque-Bera

Normality Test)..………………………………………………….……………..…...89

Appendix 3: Model Specification Test of Multiple Linear Regression Model (Ramsey

RESET Test) ………………………….………………………………………..........90

Appendix 4: Multicollinearity Testing of Multiple Linear Regression

Model……………………………………………………………………….…..……91

Appendix 5: Heteroscedasticity Testing (Autoregressive Conditional

Heteroscedasticity (ARCH) Test) …………..….…....……………………................92

Appendix 6: Autocorrelation Testing (Breusch-Godfrey LM Test)………..........…..99

Appendix 7: Residual Graph of the Multiple Linear Regression Model………...…106

Appendix 8: Empirical Result of Semi-logarithmic: Log-Lin Model….………..…107

Appendix 9: Normality Test of Semi-logarithmic: Log-Lin Model (Jarque-Bera

Normality Test)………………………………………….……………………….....108

Appendix 10: Model Specification test of Semi-Logarithmic: Log-Lin Model

(Ramsey RESET Test)…………………………………….……………..…………109

Appendix 11: Empirical Result of Model 1……………….…………….……….…110

Appendix 12: Empirical Result of Model 2……………….……………….…….…111



Appendix 15: Normality Test of Model 1 (Jarque-Bera Normality Test)….……....114


xiv

Appendix 16: Model Specification Test of Model 1 (Ramsey RESET Test)…..…..115

Appendix 17: Multicollinearity Testing of Model…………………………...……..116


Heteroscedasticity (ARCH) Test) of Model 1…………………………..……..…...117

Appendix 19: Autocorrelation Testing (Breusch-Godfrey LM Test) of Model 1….124

Appendix 20: Residual Graph of Model 1………………………….……...….........131

Appendix 21: Heteroscedasticity Problem Solving of Model 1 by using White’s

Heteroscedasticity-consistent Variances and Standard Errors Methods………..….132

Appendix 22: Empirical Result of Model 5…………………………………......….133

Appendix 23: Normality Test of Model 5 (Jarque-Bera Normality Test)....….........134

Appendix 24: Model Specification Test of Model 5 (Ramsey RESET Test)……....135


Heteroscedasticity (ARCH) Test) of Model 5……………………………..……….136

Appendix 26: Autocorrelation Testing (Breusch-Godfrey LM Test) of Model 5….143

Appendix 27: Residual Graph of Model 5…………………………………..….......150

Appendix 28: Heteroscedasticity Problem Solving of Model 5 by using White’s

Heteroscedasticity-consistent Variances and Standard Errors Methods…..………..151

Appendix 29: Hypothesis Testing Overall Significance and Individual Regression

Coefficients Significance of Multiple Regression of Model 1……………..………152

Appendix 30: Hypothesis Testing Overall Significance and Individual Regression

Coefficients Significance of Multiple Regression of Model 5………………..……156


xv

PREFACE

This research paper is submitted as a part of the requirement to fulfill for the

Bachelor of Finance (Hons) course. The title chosen for this research project is

“Factors affecting foreign direct investment decisions in Malaysia”. It revolves

around the determinants of the foreign direct investment inflows in Malaysia.

Foreign Direct Investment (FDI) is one of the key drivers in speeding up the

development and economic growth in Malaysia. Sound macroeconomic management,

presence of a well functioning financial system and sustained economic growth has

made Malaysia an attractive country for FDI. Moreover, FDI plays a crucial role in

Malaysia economy as it generates economic growth by increasing capital formation

through the expansion of production capacity.

It is reported that the charm of Malaysia in attracting FDI had declined

eventually from 1992 until 2001. It was then increased from 2002 to 2006 but

dropped significantly from 2007 to 2009. Surprisingly, FDI inflow in Malaysia

increased dramatically in 2010. The high volatility of FDI inflows to Malaysia has

drawn attention to the further study of the determinants of FDI inflow in Malaysia.


xvi

ABSTRACT

Foreign Direct Investment (FDI) plays a crucial role in speeding up the

development and economic growth of a country. In particular, developing countries

rely heavily on FDI to promote their economy as they face capital shortage for their

development process. FDI not only brings in capitals and technology, but also skills

into developing countries. And these ended up helping the countries to grow faster by

satisfying the country’s needs.

The strong growth performances experienced by Malaysia economy greatly

depends on the FDI. FDI generates economic growth by increasing capital formation

through the expansion of production capacity, promotion of export growth and

creation of employment in Malaysia. FDI inflows of Malaysia started fluctuating

from 1996 to 2010 and this high volatility of Malaysia FDI inflows drew the

researchers’ attention to examine the factors affecting FDI inflows in Malaysia by

using the annual data from year 1982-2010. Multiple linear regressions model is

applied to study the relationship between explanatory variables (market size,

economic growth, exchange rate, quality of infrastructure, trade openness, inflation

rate and China FDI inflow) and explained variable (Malaysia FDI inflow).

Empirical results show that market size, economic growth, trade openness,

inflation rate and China FDI inflow significantly and positively affect Malaysia FDI

inflows. Other than that, exchange rates also significantly affect Malaysia FDI

inflows; when Ringgit Malaysia depreciates against other currencies, FDI inflows of

Malaysia decrease. Last but not least, quality of infrastructure failed to establish a

significant relationship with Malaysia FDI inflows.


Page 1 of 157

CHAPTER 1: RESEARCH OVERVIEW

1.0 Introduction

Chapter 1 covers the brief introduction on the topic researchers chose to do in this

study, starting from the big picture narrowing down to the field for which

researchers focused on. First off, some studies are done on Foreign Direct

Investment (FDI), defining and explaining how it works. The effects FDI brought

on to a country will also be discussed in this section. Follow on, the topic are

narrowed down on to the chosen country – Malaysia. In this section, problem

statements regarding the FDI in Malaysia and the involved independent variables

will be introduced. Research objectives are then written down with more in depth

as researchers go on further into the topic. Research questions and hypothesis on

the study will also be inscribed accordingly. Under the sub topic significance of

study, the importance and contributions of the research will be discussed. Before

the chapter ends, a layout on each chapter will also be briefly outlined. Altogether,

there will be 5 chapters including: Chapter 1: Research Overview, Chapter 2:

Literature Review, Chapter 3: Methodology, Chapter 4: Data Analysis, and

Chapter 5: Discussion, Conclusion and Implications. Last but not least, Chapter 1

is concluded by providing a summary and linkage to the next chapter covering the

literature review.

1.1 Research Background

The Economy Watch (2010) defines Foreign Direct Investment (FDI) as a type of

investment involving the injection of foreign funds into an enterprise that operates

in a different country of origin from the investor. More specifically, FDI refers to

the investment of foreign asset into domestic goods and services and this does not

include the foreign investments in stock markets. FDI can be carried out through

joint ventures, Greenfield investments and cross-border acquisitions. Joint venture


Page 2 of 157

is a shared ownership with the local investors in a foreign business. This strategy

will turns out good if the MNE finds the right local partner as it can reduce

political and country risks, which in turn, increase the understanding on the local

market. However, if the wrong partner is chosen instead, political risk and agency

costs may occur. Meanwhile, a Greenfield investment is to establish a production

or service facility “starting from ground up”. It usually requires extended periods

of physical construction and organizational development. Cross-border acquisition,

on the other hand, is to directly acquire a company in the targeted country. This

requires a short period of time to gain presence and it is also a cost-effective way

of gaining competitive advantages such as brand names valued in the targeted

market (Moffett, Stonehill, & Eitheman, 2009).

According to Awan, Khan, and Zaman (2011), FDI is an essential component to

the efficient functioning of International Economic system as it speeds up the

development and economic growth of a country. However, the FDI benefits in

which host countries can expect to receive depend on the co-operation of their

government. The authors also pointed out that FDI mobilizes the capital from rich

countries to capital scarce countries. As a result, both countries can gain from this

capital movement. Shortage of capital for the development process has always

been a key problem in developing countries (Aqeel & Nishat, 2004). This is

mainly because domestically generated resources are insufficient to satisfy the

growing needs of investments in education, infrastructure and exploitation of

natural resources, thus, resulting in their inability to generate internal savings that

meet their investment needs (Vadlamannati, Tamazian, & Irala, 2009). FDI

inflows act as the lifeblood to developing countries as it brings capital to their

countries. Other than that, it made possible the transfer of technology and

managerial skills, increase in employment and enhancement in the productivity of

home country (Awan, Khan, & Zaman, 2011). Besides, FDI also benefits

investors in developed countries by enabling them to take ownership advantage in

the host country and gain profits. As a result, there is mutual benefit in the

international movement of capital among countries.


Page 3 of 157

Despite the advantages of FDI, it had also led to a few negative effects. First and

foremost is the repatriation of investment income. When foreign investors invest

in the host country, they are compensated in the form of dividends. It will then be

brought back to their country, thus, causing an outflow of fund for the host

country. The next problem is the high import content. The large inflow of FDI into

the country has brought about an increase in the imports of intermediate goods,

consequently, growth in the import bill. “Crowding-out” effects also make up as

another problem of FDI. As foreign investors invest in the host country, it

increases industry concentration and market power of a few large firms. This in

turns, create barriers for other small firms to enter (Wong & Jomo, 2005). In

conclusion, FDI brings both advantages and disadvantages to the nation‟s

economy.

Moffett, Stonehill, and Eitheman (2009) also explained that the motives of

Multinational Enterprise (MNE) investing abroad can be summarized into 5

categories comprising of market seekers, raw material seekers, production

efficiency seekers, knowledge seekers and political seekers. Market seekers

produce in foreign countries and can either export to other markets or used to

satisfy the local demands. On the other hand, raw material seekers extract raw

materials that they can find in other countries. They then either use them for

export or further processed and sell it in the country in which the raw materials are

found. Production efficiency seekers have similar concept with the raw material

seekers. They prefer to produce in countries where one or more factors of

production are underpriced in relation to their productivity. Following are the

knowledge seekers who operate in foreign countries to gain access to technology

or managerial expertise. With better technology or managerial expertise, one can

increase productivity and reduce the cost. Hence, achieving the primary objective

of MNE investing abroad – reduction of cost. Last but not least, political safety

seekers acquire or establish new operations in countries with low economy and

political risks.


Page 4 of 157

1.1.1 Foreign Direct Investment in Malaysia

Foreign direct investment is the key driver underlying the strong growth

performances experienced by the Malaysian economy. Sound

macroeconomic management, presence of a well functioning financial

system and sustained economic growth has made Malaysia an attractive

country for FDI. Other than that, the government policy reforms like

introduction of the Investment Incentives Act in 1968, establishment of

free trade zones in the early 1970, and the provision of export incentives

alongside the acceleration of open policy in the 1980s has attracted a large

amount of FDI inflow in the late 1980s (Ang, 2008). The sharp increase in

FDI of Malaysia was due to the coincidence of the foreign investment

regime which was further liberalized as part of the structural adjustment

reforms implemented in response to the macroeconomic crisis in the mid-

1980. In addition, the move by firms from Japan, South Korea, United

States and Taiwan in relocating their production bases to low-cost

countries due to the rising wages in the domestic countries also plays a part

in the increment of FDI in Malaysia (Athukorala & Waglé, 2011).

In the second half of 1980s and 1990s, the total FDI inflow into ASEAN

countries increased dramatically from an annual level of US$3 billion to

US$30 billion. Singapore remained by far the largest recipient of FDI in

the region, whereas Malaysia accounted for 25% of the total inflows into

ASEAN countries (Athukorala & Waglé, 2011). According to Karimi,

Yusop, and Law (2010), based on the result of TOPSIS method which is

used in ranking ASEAN countries in term of attraction and capacity for

FDI in 2005, Malaysia was at the second place whereas the first ranking

belongs to Singapore. This shows that Malaysia is the most attractive

country for FDI among the ASEAN countries right after Singapore.

FDI plays several crucial roles in Malaysia economy; the most crucial one

is to generate economic growth by increasing capital formation through the

expansion of production capacity. The second role is to promote export


Page 5 of 157

growth. Investing firms which have its own product reputation and brand

image in the international market reduces the need for domestic firms to

spend resources and time to penetrate and acquire foreign markets. The

facilitation of the new technology transfer to the host country and

reduction in unemployment through the expansion of the economy and job

creations resume as the third role of FDI. In addition, FDI also acts as an

agent of transformation in the Malaysian economy. This is proven with the

dominance of the influx of FDI into the manufacturing sector in its

transformation from agricultural economy to industrialized economy

(Abdul Rahim, 2006). Wong and Jomo (2005) point out that FDI can bring

in foreign exchange to be used in the payment of necessary capital and

intermediate goods imports, consequently, solving the foreign exchange

problems.

1.2 Problem Statement

As observed from Figure 1, researchers found that Malaysia‟s FDI net inflows

(BoP, current US$) were decreasing from 1992 reaching to a minimum point in

2001 with a total amount of FDI net inflows (BoP, current US$) of

US$ 553,947,368.42 only. This is the lowest amount attained since 1980s. The

average FDI net inflows (BoP, current US$) that was decreasing since 1992 was

able to increase later from 2002 to 2006. However, the FDI net inflows (BoP,

current US$) dropped significantly from US$ 8,590,185,403.74 in 2007 to

US$ 1,387,393,683.06 in 2009. Surprisingly, things turned the other way round in

2010 as the FDI inflow in Malaysia increased dramatically and reached a net

amount of US$ 9,509,265,455.11. It is by far the highest amount achieved among

the recent years. According to Athukorala and Waglé (2011), Malaysia‟s

impressive FDI inflow was severely disrupted by the financial crisis from 1997 to

1998 as they see the magnitude of FDI in Malaysia dipped during the period. The

high volatility of FDI inflows in Malaysia has drawn attention to the further study

of the determinants of FDI inflow in Malaysia.


Page 6 of 157

It is reported that the charm of Malaysia in attracting FDI had declined. As

explained by Karimi, Yusop, and Law (2010), even though Malaysia was the

second most attractive and highest capacity for foreign direct investment among

the ASEAN countries in 2005, the recent inward FDI performance of Malaysia

shows that the country was underperforming. The FDI of countries around

Malaysia like Thailand and Vietnam have surpassed Malaysia.

Figure 1: Total FDI inflows in Malaysia (BoP, current US$): 1970 - 2010

Source: World Bank (2011): World Development Indictors (Edition: April 2011).

ESDS International, University of Manchester.

The decrease of FDI in Malaysia could affect the economic growth of Malaysia

conversely. This is because previous studies showed that the strong economic

growth of Malaysia depends largely on the FDI. It injects capital and brings in

both managerial skills and technology to Malaysia with the aim of satisfying the

growing needs of domestic investment (Ang, 2008; Vadlamannati, Tamazian, &

Irala, 2009; Abdul Rahim, 2006; Athukorala & Waglé, 2011).

For better understanding of the paper, some necessary knowledge about the

independent variables is discussed briefly. First off is the independent variable -

economic growth. Slow economic growth is where the increment amount of goods

and services produced by the economy is low; this implies that the market size is

not growing rapidly and the purchasing power of the residents in the country

0.00

2,000,000,000.00

4,000,000,000.00

6,000,000,000.00

8,000,000,000.00

10,000,000,000.00

19

70

19

72

19

74

19

76

19

78

19

80

19

82

19

84

19

86

19

88

19

90

19

92

19

94

19

96

19

98

20

00

20

02

20

04

20

06

20

08

20

10

Foreign direct investment, net inflows (BoP, current US$)


Page 7 of 157

increase sluggishly. The retarded economic growth in host country discourages

investors to invest in the host country itself as it does not offer any beneficial

opportunities for investors. Foreign investors aiming at making profits prefer

growing economies to large economies (Demirhan & Masca, 2008). There is no

reason for the investors to invest in a sluggishly growing economy as the rates of

return for the investors is low and the duration to get back their principal is longer

in comparison to investing in a rapidly growing economy. This is because slow

growing economy affects the product sales, and thus, the growth of profit.

Therefore, given that the percentage of return receives year after year comes short,

investors will not be satisfied and will no longer be motivated to make any further

investment anymore. In particular, this will be in controversial with the ultimate

objective of market-seeking firm, which is to expand the business to a larger

market in order to earn more profit. In short, economic growth is an important

independent variable to be included in this study.

The next independent variable is the market size. Jordaan (as cited in Demirhan

and Masca, 2008) mentioned that FDI will move to large expanding markets with

greater purchasing power in which firms can potentially get profit from

investment. The main objective multinational enterprises expand their business

abroad is so that they can produce abroad as locals and serve the local and

regional markets without any imposition of import tax. Small market size implies

that the purchasing power and the demand of residents are low. There are not

many opportunities for foreign investors to expand business into small market as

small market size provides less efficient utilization of resources and exploitation

of economies scale. Firms always take advantage of economies of scale so that

they can produce in a larger quantity at a reduced cost. Small market size prohibits

the firm from enjoying such advantage because with the given market size and

demand of product, there is no reason for the firm to produce more than what the

market demand. As the quantity of products produced is small, the fixed cost per

unit increase. This is because fixed cost like rental and salary of employee are

invariant to the number of product produced. Thus, the lower the number of

production, the higher the average cost of product. This process is better known as

the “diseconomies of scale”. With the higher product cost, firms are unable to earn

more profit or increase their competency by setting a lower selling price. This


Page 8 of 157

creates more opportunities for producers who are capable of producing at a lower

cost and sell it at a lower price to enter the industry. The level of competition both

from and for the foreign investors has increased. Hence, if the foreign firms are

unable to compete with the other firms, the possibility of the firms coming down

to bankruptcy is high. For this reason, the risk that foreign investors have to bear

with the choice of investing in small market size is further increased. Thus, it is an

obvious fact that FDI is not in favour to be invested in small market size country.

From the explanation above, researchers can see clearly that market size is an

important factor that affects the decision of FDI.

The third independent variable is China FDI inflow which is less studied as the

factor of FDI in Malaysian case. China could be a threat to other countries nearby

like Malaysia, Thailand, Vietnam and Indonesia. According to Chantasasawat,

Fung, Iizaka, and Siu (2004), several governments have publicly noted that the

emergence of China has diverted direct investment away from their economies.

And policymakers throughout the region are convinced that the rise of China has

contributed to the foreign and domestic investors leaving their countries and

investing in China instead. China is a large country with an outstanding capability

to attract more FDI into its country than any others countries. With the high

population and market size of billions of people and the availability of large lands

for foreign investors to build their business, no wonder there is so many foreign

investors interested in the country. In other countries where there is limited land,

the price of land might be higher due to the short of supply. High population in

China creates high labour force, thus, reducing the cost of labour. Foreign

investors are attracted by these benefits and lots of the manufacturers choose to

build their factory in China to take advantage of the cheap production costs and

increased profits. With China‟s large market size and high demand, foreign

investors will be able to enjoy the benefits of economies of scale as they produce

and invest in China. Consequently, as large amount of FDI goes into China, the

neighbouring countries will only be able to shares out the remaining amount of

FDI. This leaves negative effects on the countries that heavily rely on FDI. As a

result, it is crucial to take account of China FDI inflow in the determination of

FDI inflow in Malaysia.


Page 9 of 157

Other than China FDI inflows, exchange rate is also an important factor that

affects FDI. Exchange rate is of the main concern when foreign firms make

decisions on the choice of investment because it has large impact on the capital

invested. Foreign investors do not like to invest in country with high currency

value. This is because high currency value reduces the capital of the investments.

For instance, the exchange rate of Malaysia and United States is RM3.5 / US$ 1

and RM 2.5 / US$ 1 respectively. When US firms choose to invest in Malaysia

with the amount of US$ 10, 000,000, the firm can acquire RM 35, 000,000 of

capital in Malaysian Ringgit if the exchange rate is RM 3.5 / US$ 1. However, if

the Malaysian Ringgit appreciate to RM 2.5 / US$ 1, the capital that can be used

in Malaysia to make investment is substantially reduce to RM 25,000,000. From

the above situation, researchers can see that the appreciation in the currency value

of the host country (Malaysia) reduce the capital that the investors can use to

make investment in host country. Smaller amount of fund have to be distributed

among the purchase of raw materials, hire of labour and construction of building.

As a result, high currency value is not preferable by investors. Foreign investors

like depreciated currency value because it would lead to higher relative wealth

position of foreign investors, and hence, lower relative cost of capital (Ang, 2007).

Due to the effect of the exchange rate, it is vital to include exchange rate in our

study of factors affecting FDI in Malaysia.

The fifth factor which is the inflation rate represents the stability of economy. The

higher the inflation rate, the lower the economic stability. The low inflation rates

have been effective in attracting FDI to developing country (Demirhan & Masca,

2008). The low economic stability increases the risk of the investors in face of

losses. During high inflation period, the general prices of goods and services rise.

This erodes the purchasing power of public as they need more money to buy a

product in comparison to the time period before inflation. Eventually, the quantity

of goods and services demanded will decrease. The drop in the quantity demanded

will also lead to the decrease in sales. Moreover, the cost of raw materials needed

for production increase as well and firms will not be able to exploit the advantage

of low production costs. For instance, previously RM 10,000 can purchase 1000

units of woods to produce chairs. However, with the same amount of money

during inflation period, the producer will only be able to purchase 800 units of


Page 10 of 157

wood. As the average cost of production rise, the selling prices of the product

increase, leading the public to the inability to meet up with the expenses. And in

the end, it will negatively affect the profit of the business and indirectly affects the

return of the investors. As a result, due to the impact inflation rate have on the

profitability of business, it is important to be considered by investors before

making any investment in that country.

Other than that, infrastructure quality is also another determinant of FDI.

Infrastructures such as road, ports, railways and telecommunication system are the

basic needs of firms in support of daily business routine. Poor infrastructures that

reduce productivity and potential of investments are major constraint for low-

income countries. Cost of transport and delivery time will be increased due to the

poor infrastructure. OECD points out that although a lot of interest arose among

foreign investors on the country of China emerged after 1979, large FDI inflow

did not occur in the initial period due to the poor infrastructure (as cited in Ali and

Guo, 2005). As market-seeking firms invest and produce in foreign countries, it

will have to deliver it to different regions of the countries after production so as to

serve the local consumers‟ needs. Poor road and railway condition would increase

the possibility of transport break down, delay in delivery time and damage of

products on the way to its destination. Multinational enterprise that set up

subsidiary in host country takes telecommunication seriously because of its role as

a bridge that connects both the parent and subsidiary company. Poor

telecommunication services such as problematic internet connection restraint

parent company from doing distance monitoring and supervising the activities of

their subsidiary. Nowadays, company use video conferencing to monitor their

subsidiary and conduct meeting with other company in order to save time and cost.

For all these reasons, researchers are convinced that infrastructure quality is

significant in attracting FDI.

Last but not least is the trade openness of the host country. Trade openness

indicates the degree host country response to trade, and it involves both import

and export activities. Country which dislikes import and export would impose a

high tariff on both imported and exported goods. This would discourage foreign

investors to make investment in host country, in particular, export-oriented firms.


Page 11 of 157

According to Aqeel and Nishat (2005), horizontal FDI is motivated by lower trade

cost, hence, high tariff barriers induce firm to engage in horizontal FDI to replace

exports with production abroad by foreign affiliates. Export-oriented firms import

materials that cannot be found in host country to be further processed and

exported to other countries for sale. So if high tariff is imposed on imported goods,

the cost of producing the product would increase and the volume of import will

decrease. This will harm both the productivity and the profitability of the firm.

Vertical FDI can be characterized by individual affiliates specializing in different

stages of production output and semi products, which in turns, are exported to

other affiliates for further processing (Aqeel & Nishat, 2005). By using this

fragmenting production process, it enables company to take on different cost

advantages at different countries. For instance, ABC Company faces the problem

of high assembling cost and less profitable sales in country A but yet do not want

to give up on their business. Fortunately, the cost of labour in country B is very

low. Company A can opt to reduce their cost significantly with the assembly of

products done in country B and then export back to country A for sale. Therefore,

by acquiring the material at a lower cost in country A and assembling the final

product in country B, the cost of the product is reduced in comparison to it is

finished in either of the country. This fragmentation process gives ABC Company

an opportunity to invest abroad and reduce the cost of production. However, given

that the trade openness in country B is low with the imposition of high tariff on

import and export product, the cost after the taxes will be much higher than before

the fragmentation process. The imposed taxes in country B have given up the

chance to attract foreign direct investment into country B.

In conclusion, the seven factors which are made up of economic growth, market

size, China FDI inflow, exchange rate, inflation rate, infrastructure quality and

trade openness are important in the decision making of FDI. Therefore,

researchers have included them in this study to verify whether there is relationship

between these factors and FDI in Malaysia.


Page 12 of 157

1.3 Research Objectives

1.3.1 General Objective

The general objective is to examine the relationship between FDI inflows

and the independent variables in Malaysia from 1982-2010.

1.3.2 Specific Objectives

i. To examine the relationship between economic growth and FDI

inflows in Malaysia from 1982-2010.

ii. To examine the relationship between market size and FDI inflows

in Malaysia from 1982-2010.

iii. To examine the relationship between China FDI inflows and

Malaysia FDI inflows 1982-2010.

iv. To examine the relationship between exchange rate and FDI


v. To examine the relationship between inflation rate and FDI inflows

in Malaysia from 1982-2010.

vi. To examine the relationship between quality of infrastructure and

FDI inflows in Malaysia from 1982-2010.

vii. To examine the relationship between trade openness and FDI


1.4 Research Questions

i. Is there any significant relationship between FDI inflows and at

least one of the independent variables in Malaysia from year 1982-

2010?

ii. Is there any significant relationship between economic growth and

FDI inflows in Malaysia from year 1982-2010?


Page 13 of 157

iii. Is there any significant relationship between market size and FDI

inflows in Malaysia from year 1982-2010?

iv. Is there any significant relationship between China FDI inflows and

Malaysia FDI inflows from year 1982-2010?

v. Is there any significant relationship between exchange rate and FDI


vi. Is there any significant relationship between inflation rate and FDI


vii. Is there any significant relationship between quality of infrastructure

and FDI inflows in Malaysia from year 1982-2010?

viii. Is there any significant relationship between trade openness and FDI


1.5 Hypotheses of the Study

H0: There is no relationship between all independent variables and FDI

inflow in Malaysia.

H1: At least one independent variable has relationship with FDI inflow in

Malaysia

H0: There is no relationship between economic growth and FDI inflow in

Malaysia.

H1: There is relationship between economic growth and FDI inflow in

Malaysia.

H0: There is no relationship between market size and FDI inflow in

Malaysia.

H1: There is relationship between market size and FDI inflow in Malaysia.

H0: There is no relationship between China FDI inflows and FDI inflow in

Malaysia.

H1: There is relationship between China FDI and FDI inflow in Malaysia.

H0: There is no relationship between exchange rate and FDI inflow in

Malaysia.


Page 14 of 157

H1: There is relationship between exchange rate and FDI inflow in

Malaysia.

H0: There is no relationship between inflation rate and FDI inflow in

Malaysia.

H1: There is relationship between inflation rate and FDI inflow in Malaysia.

H0: There is no relationship between quality of infrastructures and FDI

inflow in Malaysia.

H1: There is relationship between quality of infrastructures and FDI inflow

in Malaysia.

H0: There is no relationship between trade openness and FDI inflow in

Malaysia.

H1: There is relationship between trade openness and FDI inflow in

Malaysia.

1.6 Significance of the Study

Determinant of FDI is a popular topic among the researchers. Even though, there

have been many previous studies done on the determinants of FDI in Malaysia, in

this case, researchers have added a relatively new variables - FDI inflow of China

- into the model in order to find out whether the amount of FDI inflow to China

affects the FDI inflow of Malaysia. There have not been many researches that

included China FDI inflows as an independent variable in the examination of the

determinants of FDI inflow in Malaysia. Other than that, researchers form a new

conceptual model which differs from previous studies. Researchers modify the

theoretical framework by picking out the factors they are interested in examining

and also adding in a new variable, China FDI inflows.

This study will contributes to policymakers like Bank Negara Malaysia and the

Federal Government as it gives them a picture of what variables are significantly

affecting FDI inflows in Malaysia. Researchers have included some important

economic factors like economic growth, market size, exchange rate, inflation rate,

quality of infrastructures and trade openness. The most important factors are of


Page 15 of 157

course the FDI inflows of China. Bank Negara Malaysia and Federal Government

play an important role in affecting Malaysia‟s economic environment through the

monetary policy and fiscal policy. Monetary policy is used by Bank Negara

Malaysia to stimulate economic by controlling both the money supply and

demand. On the other hand, fiscal policy is where the government uses the

expenditure and revenue (taxes) to influence the economy. This study results can

serve as a guideline or reference to Bank Negara Malaysia and the Federal

Government in formulating monetary and fiscal policy to meet up with the

preference of direct investors who consider investing in Malaysia. Besides, these

can prevent policymakers from focusing on the unnecessary areas wasting

resources in an effort to attract more FDI. With the huge amount of FDI, it will be

able to stimulate Malaysia‟s growth, increase employment rate, living standards

and technology transfer and also shorten the period to achieve Vision 2020.

China is a large country with low labour cost, large market size, and high

productivity level. For all these reasons, China easily out wins other countries in

attracting a much higher FDI into its own country. Most of the manufacturing

firms choose to invest in China to exploit the cost advantage. If in this study,

researchers found out that FDI inflow of China has significant negative

relationship with the amounts of FDI inflows in Malaysia, then the Federal

Government of Malaysia should avoid direct competition with China. In contrast,

if it is found that there is a significant positive relationship between the both,

Malaysia should maintain a good relationship with China. They may consider

improving the trading transaction with China or may be even come together with

China in constructing policy which benefits both Malaysia‟s and China‟s economy.

Other than the contribution to Bank Negara Malaysia and the Federal Government,

this study also provides guidelines or serves as a reference to potential direct

investors who wish to invest in Malaysia. Before direct investors decide on

investing in Malaysia, they will perform a series of examination on Malaysia‟s

situation to determine whether or not it is profitable for them to invest in. This

study will guide them through the determinants which have significant effect on

the FDI inflows of Malaysia. It will also prevent potential direct investors from

investing in countries with high risks and negative return.


Page 16 of 157

In short, by conducting this study, researchers are able to understand more about

the determinants of FDI in Malaysia and provide a more robust result to Bank

Negara Malaysia and potential direct investors on the impact of economic growth,

market size, China FDI inflows, exchange rate, inflation rate, infrastructure

quality and trade openness have on Malaysia‟s FDI in flow.

1.7 Chapter Layout

1.7.1 Chapter 1

Chapter 1 discusses about the topic that researchers are interested to study,

introduce the topic and write out the problem statement. Other than that,

researchers will also be going through on the objective of conducting this

study, what researchers are going to investigate and also the contributions

and the importance of the study.

1.7.2 Chapter 2

Chapter 2 is the literature review part. Researchers will be summarizing on

what they understand as they read through the past researchers‟ work. This

increases researchers‟ understanding on the topic that researchers are going

to do. Besides, researchers will also review on any relevant theoretical

models and come out with the conceptual framework for the research.


Page 17 of 157

1.7.3 Chapter 3

Chapter 3 is the methodology part in which there will be a description on

how the research is carried out in term of design, data collection methods,

sampling design, operational definitions of constructs, measurement scales,

and methods of data analysis. It mainly discusses the preparation work

before moving on to the data analysis part which constitutes the next

chapter.

1.7.4 Chapter 4

Chapter 4 presents pattern of result using the data and methods previously

described in chapter 3. Then, researchers will analyse the results to answer

the research questions and hypothesis written down in Chapter 1.

1.7.5 Chapter 5

Chapter 5 is the last chapter of the research in which there will be

discussion, conclusion and implications. It summarizes the whole study

and converse the major finding, what can be recommended to policy

makers and practitioners from the result obtained in the research. Other

than that, it also point out the limitations of the study and provide

recommendation so that next researcher can further the study if he / she is

interested.


Page 18 of 157

1.8 Conclusion

This research paper introduce FDI in details with its definition, types of FDI,

motives of FDI, advantages and disadvantages of FDI and also how FDI works.

Moreover, this study discuss on the seven determinants of FDI that researchers are

interested in, including economic growth, market size, China FDI inflows,

exchange rate, inflation rate, infrastructure quality and trade openness.

Researchers also explained on this research‟s objectives - to understand the

determinants of FDI inflow in Malaysia in order to improve the future

performance of FDI inflow. In terms of the contribution, researchers hope this

research will provide policymakers with a better understanding of the factors

affecting FDI so that an appropriate policy can be developed. Other than that,

researchers also explain on the chapter layout of this study. After clearing up on

what need to do in this research, researchers proceed to the next chapter which is

the literature review. This research paper will also study on the past researchers‟

work on the relationship between FDI inflows and the seven determinants and

summaries it under the next chapter.


Page 19 of 157

CHAPTER 2: LITERATURE REVIEW

2.0 Introduction

Past research studies relating to the determinants of FDI inflows in this research

paper are summarized in this chapter. This provides a better understanding of the

nature of FDI, economic growth, market size, China FDI inflows, exchange rate,

inflation rate, quality of infrastructure and trade openness. Other than that, the

relationships between the dependent variable and independent variables are

studied as well. With the help of the previous studied models, researchers are able

to formulate a new proposed framework for this study.

2.1 Review of the Literature

2.1.1 Foreign Direct Investments

According to Moffett, Stonehill, and Eitheman (2009), foreign direct

investment (FDI) is investment undertaken by an entity resident of one

economy in an enterprise resident in another economy, with the objectives

of obtaining and sustaining a lasting interest (profits) in the enterprise and

also to exercise a significant degree of influence in its management.

Management and voting rights are granted to the investors if the investors‟

ownership level is greater than or equal to 10% of the ordinary shares.

Shares ownership less than 10% is termed portfolio investment and is not

categorized as FDI (Economy Watch, 2010). FDI can be classified into

inward FDI and outward FDI, depending on the direction of the flow of the

money. Inward FDI occurs when foreign capital is invested in local

resources whereas outward FDI refers to local resources invested in

foreign country, also named as “direct investment abroad”.


Page 20 of 157

According to Vadlamannati, Tamazian, and Irala (2009), the determinants

of FDI are divided into macroeconomic factors, institutional factors,

political factors and socioeconomic factors. Macroeconomic factors

include labor and potential macroeconomic risk like inflation rate and

unemployment rate. Institutional factors are track record of government,

corruption and civil liberties whereas political factors include political

regime and political instability. Lastly, socioeconomic factor uses literacy,

infant death and infant mortality rate as indicators. Infrastructure quality

can be considered as a factor that affects foreign direct investment as well.

According to OECD (as cited in Ali & Guo, 2005), although there was

emerging interest among foreign investors in China after 1979, large FDI

inflows did not happen in the initial period due to poor infrastructure.

2.1.2 Economic Growth

The relationship between Foreign Direct Investment (FDI) and economic

growth has been a topical issue for several decades. Many researchers have

conducted studies to investigate the positive causal relationship between

FDI and economic growth, either in the short run, or in the long run, or

both. On top of recognizing the importance of FDI to growth, economic

growth itself has also been identified frequently as an important

determinant of FDI inflow into the host countries (Benacek, Gronicki,

Holland, & Sass, 2000).

According to Hansen and Rand (2006), rapid growth of an economy might

attract more FDI by multi-national companies (MNCs) as they locate new

profit opportunities. Dunning (1995) argued that MNCs with certain

ownership advantages will invest in another country with locational

advantages, and both advantages can be captured effectively by

“internalizing” production through FDI. This market hypothesis has been

tested in many empirical papers (Chakraborty & Basu, 2002; Moosa,


Page 21 of 157

2002). High GDP growth rate represents soundness and stability of

economic policies, and the effectiveness of the government institutions

which are mainly looked for in international transactions. Thus, it will

cause the levels of aggregate demand for investments (both domestic and

foreign) to rise (Zhang, 2001).

Fan, Morck, Xu, and Yeung (2007) encouraged by past growth

performance, also note that foreign investors overflow China in

anticipation of improved institutions. With the aid of panel data for 80

developed and developing countries, Choe (2003) conducted a Granger

causality test for GDP and FDI. It is found that the causality between

economic growth and FDI runs in either direction but with a tendency

towards growth causing FDI; there is little evidence that FDI causes host

country‟s growth. Thus, a significant and positive relation is once again

proven between GDP growth rate and FDI inflows in a country.

While most studies found the importance of economic growth on FDI,

there are also other studies which failed to validate the hypothesis. For

instance Kahai (2011), on the other hand, could not established a

significant relationship between economic growth (measured as the annual

real GDP growth rate) and FDI using the data from 1998 and 2000 for

fifty-five developing countries. Although all these studies provide ample

evidence of the relationship between economic growth and FDI in both

developed and developing countries, few studies have been done in the

case of Malaysia. Therefore, economic growth is included as one of the

independent variables in this research paper.

2.1.3 Market Size

Market size has been proved to be one of the most important determinants

of FDI by numerous past empirical studies (Lim, 2008; Luiz &

Charalambous, 2009; Ang, 2008; Athukorala & Waglé, 2011).Market size


Page 22 of 157

of a country represents the potential demand for the country‟s output and

also its economic conditions. It is an important element that will determine

the foreign direct investors‟ investment in a particular country (Asiedu,

2002). For those countries which have large markets, the stock of FDI is

expected to be larger than those of the small markets‟. Market size is

normally measured by real GDP or GDP per capita GNP. At times, private

and public consumption are also used as alternatives.

Majority of studies use GDP and GDP per capita as a proxy for market

size and it is found that there is a positive relationship between market size

and FDI inflow to the country (Artige & Nicolini, 2005). According to

Sharma & Bandara (2010), investors are easily attracted to large expanding

market. Although there still remains other factors that might affect foreign

investors‟ decision making, the first things to make them have the

intention to invest in a country is none other than the size of market (Awan,

Khan, & Zaman, 2011). This is because market that is small and unable to

expand rapidly does not possess any inherent attractiveness. Charkrabarti

(2001) (as cited in Moosa & Cardak, 2006) stated that a larger market size

of a country indicates that the country will be more efficient in utilizing

their resources and exploitation of economic of scale. Hence, small market

size country will lose its competitiveness in comparison to such countries

in attracting more investors (Medvedev, 2012).

As mentioned above, it is necessary to consider market size as an

important factor in determining FDI inflows in a country (Asiedu, 2002).

However, at the same time, it is not the only factor influencing FDI.

Medvedev (2012) argued that the barrier of trade in a country will affect

the FDI inflow to the country even when the market size is large. On the

other hand, Nurudeen, Wafure, and Auta (2011) and Bevan and Estrin

(2004) found that GDP have significant but negative effect on the FDI.

Despite the increasing country size, foreign investors are less willing to

invest in a particular country which they have less perceivability on the

economy (Nurudeen, Wafure, & Auta, 2011).


Page 23 of 157

Also, some studies found that GDP is not suitable to be used as a proxy for

market size. According to Demirhan and Masca (2008), the empirical

results showed GDP insignificant to FDI because absolute GDP reflects

the size of population rather than the income. It is also suggested that GDP

growth rate and growth per capita GDP will be a more suitable proxy of

market size. However, this argument has been less supported by other

researchers.

Many researchers have proved through empirical studies for market size

that using GDP or GDP per capita showed significant and positive

correlation with FDI inflow to both developing and developed countries

(Quer & Claver, 2007; Rodriguez & Pallas, 2008; Vadlamannati,

Tamazian & Irala, 2009; Trevino & Mixon Jr, 2004). Therefore, in this

study, researchers will be using GDP or GDP per capita as a proxy to

market size and proved on the significant positive relationship it has with

the FDI in Malaysia.

2.1.4 Inflation Rate

Inflation rate is taken as a proxy for the level of macroeconomic stability

of a country. Usually, high rate of inflation, so called the unbridled

inflation, in a country will reduce the return on investment and act as an

indicator of macroeconomic instability. It is considered as a sign of

economic tension and unwillingness of the government to balance its

budget and failure of the central back to conduct appropriate monetary

policy (Azam, 2010). A low inflation rate is taken as a sign of internal

economic stability in the host country, reflecting a lesser degree of

uncertainty which encourage foreign direct investment (Asiedu, 2002). In

short, there is a negative relationship between inflation rate and foreign

direct investment. Demirhan and Masca (2008) did a research on the

determining factors of foreign direct investment inflow in 38 developing

countries over the period of 2000-2004 by estimating a cross-sectional


Page 24 of 157

econometric model. According to the econometric results of this past

research, it is found that inflation rate has a negative sign and is

statistically significant to foreign direct investment. It means that low

inflation rates have been effective in attracting foreign direct investment

into developing countries. Besides that, other researchers like Azam (2010)

and Shamsuddin (1994) also found that there is a significant negative

relationship between inflation rate and foreign direct investment.

On the other hands, some researchers think that there might be a positive

relationship between inflation rate and foreign direct investment.

Srinivasan (2011) states that higher inflation indicates higher price levels

and increased in the production activities of the host country and attraction

of investments from foreign firms, which then leads to an increased

expected level of profitability. In this research paper, Srinivasan (2011)

used fixed effects and random effects models to explore the determinants

of foreign direct investment in the selected South Asian Association for

Regional Cooperation (SAARC) countries for the period of 1970-2007.

This paper showed that inflation rate is one of the most significant factors

in determining foreign direct investment in SAARC countries. However,

there is some past research papers that indicated inflation rate is

insignificant to foreign direct investment (Vijayakumar, Sridharan & Rao,

2010; Nurudeen & Wafure, 2010). Nurudeen, Wafure, and Auta (2011)

examined the major determinants of foreign direct investment in Nigeria

by analyzing the annual data over the period 1970 – 2008. Using the

ordinary least squares and error correction techniques, the regression

results showed that inflation rate is insignificant but have positive

influence on the foreign direct investment inflows.

In a nutshell, based on the past researches, inflation rate may have both

significant or insignificant and negative or positive influence on foreign

direct investment.


Page 25 of 157

2.1.5 China FDI Inflow

China FDI inflow is the main variable that we will be examining in this

paper. There has been a few past research papers determining the influence

of China FDI inflow on the foreign direct investment. Most of the past

research papers showed that there will be a significant relationship

between China FDI inflow and foreign direct investment either in a

positive or negative relation (Salike, 2010; Chantasasawat, Fung, Iizaka, &

Siu, 2004; Eichengreen & Tong, 2007).

According to past researchers, two types of effects will most probably be

incurred in the event. The first one is known as the investment-diversion

effect. In choosing between China and other Asian countries, multinational

enterprises may consider a host of factors including wage rates, political

risks and infrastructure that would make a particular destination desirable

as the site for low-cost production. As proven, China‟s labour cost which

is low may lure multinational enterprises away from sites in other

developing countries like Thailand during the consideration of an

alternative location for low-cost export platforms. Investing in China will

then reduce the FDI in other countries. Therefore, it will be in a negative

sign and so called the investment-diversion effect.

Second effect is the investment creation effect. The production and

resources linkages between China and other countries takes place in the

form of further fragmentation and specialization of production process.

This linkage takes advantage of the respective competitiveness of different

economies in the distinct stages of production. When components and

parts are traded between China and another economy, an increase in

China‟s foreign direct investment will be positively related to an increase

in the other economy‟s foreign direct investment. Another complementary

argument is that as China‟s economy grows, its market increases and its

appetite for minerals and resources rise too. This will lead to the

investment by other multinational enterprises in some other countries to


Page 26 of 157

extract minerals and resources in order to export it to China in need of the

whole spectrum of raw materials. In this case, there will be a positive sign

of China FDI inflow or the so called investment creation effect.

Athukorala and Waglé (2011) have examined patterns and determinants of

foreign direct investment in Malaysia from a comparative Southeast Asian

perspective. In that research paper, foreign direct investment flows into

China is taken as an additional explanatory variable to test whether foreign

direct investment in ASEAN is crowded out by foreign direct investment

into China, which has become more attractive for foreign direct investment

among the developing countries in recent years. The result of this paper

showed that there is no evidence that foreign direct investment in

Southeast Asian countries is crowded out by the increasing flow of foreign

direct investment into China. On the contrary, Malaysia (or Southeast

Asian countries) benefits from a complementary foreign direct investment

relationship with China as it becomes a favored location for high-end tasks

within the global production networks.

Eichengreen and Tong (2006) employ a gravity model to examine the

impact China have on the exports and foreign direct investment receipts of

other countries. It was asked whether there are grounds for “fear of China”.

This research showed that China‟s emergence has very different effects on

different groups of countries. They found that there is a complementarity

between inflows of foreign direct investment in China and those into other

Asian countries, but substitutability for those in Organisation for

Economic Co-operation and Development (OECD) countries. Wang, Wei

and Liu (2007) used data from a longer interval of 1980 to 2003 in

assessing the China‟s effect on individual economies. They found that

there is a significant foreign direct investment creation effect on India and

Philippines but a significant foreign direct investment diversion effect on

Indonesia, Korea, Malaysia and Taiwan.


Page 27 of 157

As a result, theoretically, the net effect of investment creation and

investment diversion for China cannot be determined prior and must be

examined empirically.

2.1.6 Exchange Rate

The effect of exchange rate volatility on FDI movement is also a fairly

well studied topic. Froot and Stein (1991), Klein and Rosengren (1994),

Guo and Trivedi (2002) and Kiyota and Urata (2004) found that

depreciation of exchange rate in the host country will in turns increase FDI

of the host country. Conversely, when the host country‟s exchange rate

appreciates, the FDI in that particular country will decrease. Nurudeen,

Wafure and Auta (2011) had used the Ordinary Least Squares and Error

Correction Techniques to study the relationship between FDI and

exchange rate depreciation. Their finding was found to be in line with the

research by Hara and Razafimahefa (2005) in which the exchange rate

depreciation significantly and positively affects FDI inflows. This means

that when a country‟s exchange rate depreciates, it will attract the foreign

investors to invest in the country as it has a lower dollar price in its

domestic industries. Besides, investors are more likely to invest in the

market when the targeted market has a weak currency. They will postpone

their investment to the period when the currency depreciates because they

believe that they will earn a higher profit in their investment at a later date.

This shows that there is a significant time lag between exchange rate

volatility and FDI movement (Barrell & Pain, 1996).

However, Aqeel and Nishat (2004) had found a controversial result. They

applied Cointegration and Error Correction Techniques in their study to

identify the variables in explaining the relationship between exchange rate

and FDI. With the average annual exchange rate of the host country as the

indicator, it is found that there is a significant and positive relationship

between the both. Hence, it is indicated that when the currency appreciates,


Page 28 of 157

FDI increases too. This is simply because investors have high expectation

on the economy and the returns. Study by Campa (1993) in US also found

that an appreciation of exchange rate in the host country will in fact

increase the FDI in the host country. This is because investors believe that

an appreciation of exchange rate in the host country will most likely

increase the future profitability in terms of the home currency.

In explaining the case of South Asian Association for Regional

Cooperation (SAARC) countries, Srinivasan (2011) found that the real

exchange rate are insignificant and in fact has a negative relationship with

the FDI. It seems that real exchange rate do not play a significant role in

attracting FDI in SAARC countries.

2.1.7 Trade Openness

The competition for inward FDI in many developing countries are

increased due to the ongoing process of integration of the world economy

and liberalization of the economies, the controls and restrictions over the

entry and operations of foreign firm which are now being replaced by

selective policies aiming at FDI inflows (Aqeel & Nishat, 2005). The

liberalization of the economies always refers to the openness of the

economy or trade and it is one of the common variables in explaining the

FDI inflows for a country. Normally, it is measured by the share of exports

and imports in GDP. According to Moosa and Cardak (2006), Demirhan

and Masca (2008), and Sawkut, Boopen, Taruna, and Vinesh (2009), trade

openness is significant and has a positive effect on FDI inflows. A

country‟s willingness to accept foreign investment is important to the FDI

of the country. In the research of Chantasasawat, Iizaka and Siu (2010), the

result is found to be the same with the previous study which show positive

significance. The study also interestingly stated that the openness of a

country includes the degree of both tariff and nontariff measures. The

reductions in different types of trade barriers will in turns increase FDI in a


Page 29 of 157

country. Awan, Khan, and Zaman (2011) also studied the relationship

between the degree of trade openness and FDI by using Augmented-

Dickey Fuller (ADF) test and Co-integration and Error Correction Model

(ECM) in their research. They took the sum of exports and imports each

year as the indicator of trade openness. The result showed that the degree

of trade openness is highly significant with a positive sign in both ADF

test and ECM. This means that the foreign investors would prefer making

investment in countries with a higher degree of trade openness. A few

studies applied the Fixed Effects (FE) model and the Random Effects (RE)

model to examine the effect of trade openness on FDI. In Srinivasan (2011)

case, it is found that the openness of trade is positive and statistically

significant to the FDI inflows. Thus, it is an obvious fact that investors will

more likely make investment in those countries which have opened up to

the outside world.

However, Kolstad and Villanger (2008) (as cited in Awan, Khan, &

Zaman, 2011) found that trade openness is insignificant in explaining the

inflows of FDI. Busse and Hefeker (2007) confirm this and add that trade

openness negatively affects FDI inflows. According to Goodspeed,

Martinez-Vazquez, and Zhang (2006), the effect of trade openness on FDI

is inconclusive because thus far different studies show different results.

Sometimes, trade openness is significant and has a positive relationship

with FDI inflows but at times it is insignificant with the other conditions of

the empirical model.

In the Organization for Economic Co-operation and Development (OECD)

countries, trade openness and FDI inflows have displayed a positive

relationship. On the other hand, small effects have been reported on non-

OECD countries and resource-rich fuel exporting countries‟ FDI inflows

(Seim, 2009). Last but not least, the author found trade openness and FDI

inflows have negative relationship in the transition countries.


Page 30 of 157

2.1.8 Quality of Infrastructure

Amongst huge literature on FDI, few scholars have actually acknowledged

the significant contribution of infrastructure in stimulating FDI inflows.

The few proponents featured are Wheeler and Mody (1992), Asiedu (2002),

Quazi (2007), Kahai (2011), and Rehman, Ilyas, Alam, and Akram (2011).

These authors have argued that good infrastructure is a necessary condition

for foreign investors to operate successfully. Poor infrastructure or

unavailable public inputs increase costs for firms. Thus to the extent that

the public input is non-excludable and non-congestible, it will lower the

costs of doing business for multinational and indigenous firms alike.

In their influential paper, Wheeler and Mody (1992) employed a translog

specification and uses a panel of 42 countries for the period 1982-1988 to

analyze the determinants of FDI. They interestingly reported that

infrastructure quality (quality of transport, communications, energy

infrastructure and degree of industrialization) exhibit a high degree of

statistical significance and thus have large, positive impacts (1.57 to 2.54)

on investment. In addition, Asiedu (2002) who analyzed 34 countries in

Africa over the period 1980-2000 using the number of telephones per 1000

population (as a measure of infrastructure development) and at the same

time controlling for the classical FDI determinants supported it and

concluded that countries that improved their infrastructure were “rewarded”

with more investments.

Meanwhile, Rehman, Ilyas, Alam, and Akram (2011) also constructed an

indicator for infrastructure that encompassed number of telephones

available per 1,000 people as measures. By using time series data from

1975 to 2008 and applying autoregressive distributive lag (ARDL)

approach to cointegration, it is concluded that there are significant positive

impact in both the short and long run of infrastructure on FDI inflows in

Pakistan. In short run, one percent increase in infrastructure results in

uplifting FDI of 1.03% and in long run, one percent rise in infrastructure


Page 31 of 157

enhances FDI inflows by 1.31%.This further proves that infrastructure has

a significant attractiveness for FDI inflows in developing economies

(Asiedu, 2002; Kahai, 2011; Mengistu & Adhikary, 2011).

While most studies found the importance of infrastructure for FDI, there

are also other studies which failed to validate the hypothesis. For instance

Quazi (2007), on the other hand, could not established positive and

significant relationship between infrastructure ( measured as the number of

telephones per 1,000 people) and FDI using panel data from 1995-2000 for

a sample of seven East Asian countries. The authors however admitted that

„it is plausible that their proxy variables - the natural log of the number of

telephones available per 1,000 people and the adult literacy rates,

respectively, perhaps inadequately capture their true effects on FDI.

Moreover, Addison et. Al. (2006) as cited in Rehman, Ilyas, Alam, and

Akram (2011) acknowledge such promotional impact only for developed

nations but, on the other hand, such situation does not exist for developing

countries.

A review of literature suggests that while the role of infrastructure in

attracting FDI has received increasing interest from academic scholars

lately, yet these studies focused on the general level of infrastructure and

moreover have largely ignored developing country cases, particularly

Malaysia economies. Thus, the current study attempts to fill in this gap and

thus supplements the growing literature on FDI.


Page 32 of 157

2.2 Review of Relevant Theoretical Models

Figure 2.1: Theoretical Model

Source: Ang, J.B. (2008).Determinants of foreign direct investment in Malaysia.

Journal of Policy Modeling, 30, 185-189.

FDIt = β0 + β1FDt + β2GDPt + β3GROt + β4INFt + β5OPEt + β6RERt + β7TAXt +

β8UNCt + β9D97-98

The chart above (Figure 2.1) showed the theoretical model by Ang (2008). This

study explore the effect of financial development, market size, economic growth,

infrastructure development, trade openness, real exchange rate, and statutory

corporate tax rate in Malaysia towards foreign direct investment inflows in

Malaysia. The study examines the determinants of FDI for Malaysia to enlighten

analytical and policy debates.

As seen from the model of Ang (2008), the independent variables involved are

financial development, gross domestic product, annual growth of gross domestic

product, infrastructure development, trade openness, real exchange rate, and

statutory corporate tax rate in Malaysia. On the other hand, the dependent variable

is foreign direct investment inflows in Malaysia.

FDIt

FD

UNC

D97-98

RER

GRO

TAX

GDP

INF

OPE


Page 33 of 157

According to Ang (2008), those factors are the key forces that stimulate FDI

inflows in Malaysia. Financial development (FDt) is measured by private credit to

GDP while gross domestic product (GDPt) is used as an indicator to determine

market size of country. GROt, which is the annual growth rate of GDP, measured

the economic growth of country. Meanwhile, infrastructure development (INFt) is

proxy by total government spending on transport and communication whereas

trade openness (OPEt) is defined as the sum of exports and imports over GDP.

RERt is the real exchange rate and TAXt is the statutory corporate tax rate in

Malaysia. UNCt represents the macroeconomic uncertainty related to output

fluctuations. Last but not least, Ang (2008) include D97-98 which refers to dummy

variable accounting for the Asian financial crisis in 1997-1998.

2.3 Proposed Theoretical Framework

Figure 2.2: Researcher‟s Model

Adapted from: Ang, J.B. (2008).Determinants of foreign direct investment in

Malaysia. Journal of Policy Modeling, 30, 185-189.

MFDI

GDP GDPG

INF TL

CFDI TO

OREER


Page 34 of 157

Based on the model from Ang (2008), we remodeled the model into (Figure 2.2):

MFDIt = β0 + β1GDPt + β2GDPGt + β3OREERt + β4TLt + β5TOt + β6INFt +

β7CFDIt

As seen from the model above, the independent variables we used as determinant

of FDI inflows in Malaysia are market size, economic growth, infrastructure, trade

openness, real exchange rate, China FDI, and inflation rate in a country. The

dependent variable is the FDI inflows in Malaysia (MFDI).

The independent variable GDP, which were initially symbolized as GDP in the

theoretical model, is the indicator for market size. According to Chakraborty and

Basu (2002), larger market size will induce more foreign investment into the

country. Hence, GDP are expected to have positive relationship with the foreign

direct investment inflows in Malaysia (Artige & Nicolini, 2005).

Independent variable GDPG is the indicator for economic growth which is

commonly proxy by annual growth domestic product growth rate. Based on

Globerman and Shapiro (2003), higher gross domestic product growth rate

indicates that the country is doing well in development. Ang (2008) proved that

growth domestic product growth rate is important and has positive impact on FDI

inflows. Thus, higher gross domestic product growth rate tends to attract more

foreign investor to invest in that particular country.

The independent variable OREER refers to the indicator for official real exchange

rate. Hara and Razafimahefa (2005) stated that when the exchange rate of a

country depreciates, it will attract more FDI inflows. Theoretically, it is true

because foreigners tend to invest in lower capital countries as they wish to

generate higher income when the exchange rate of the country they invest in

appreciate in the possible future.

Independent variable TL is the indicator for quality of infrastructure. Wheeler and

Mody (1992), Loree and Guisinger (1995), Morisset (2000), and Asiedu (2002)

stated that infrastructure is one of the key indicators in determining the FDI


Page 35 of 157

inflows of a country. Good infrastructure such as increased number of telephone

line in the country allows firms to operate or promote their businesses more easily.

This in turns, increase the opportunity of foreign investor investing in the country.

Independent variable TO represents the indicator for trade openness. Theoretically,

open economies will generally generate greater market opportunity. According to

Moosa and Cardak (2006) and Demirhan and Masca (2008), trade openness is

crucial to a country‟s foreign direct investment. This is because most of the

investors prefer investing in the country which has less trade barriers.

Independent variable INF is the indicator for inflation rate of the country. In

theory, the higher the inflation rate, the more expensive the cost of domestic

product become. This will then decrease the foreigners investing in the country

due to their need to take out more cost of capital to buy resources or hire worker

in that country. Demirhan and Masca (2008) also found that inflation rate has a

significant but negative relationship with foreign direct investment.

Last but not least, independent variable CFDI is the indicator for China FDI

inflows. This variable is used to determine whether China‟s increased in its FDI

inflows will affect Malaysia FDI inflows. Based on Wang, Wei and Liu (2007),

China FDI is significant and will affect other countries‟ FDI. This is because

China imposed lower labour cost compare to other countries, so foreign investors

will most probably shift their interest and invest in China due to the need of lower

capital to operate the firms. It is predicted that an increase in China FDI will lead

to a decrease in Malaysia FDI.

2.4 Hypotheses Development

After forming a conceptual framework based on the past researcher‟s reviews,

researchers have also formed hypotheses on the model and also each variable to

examine whether the theory formulated is valid or not. Below are the hypotheses:


Page 36 of 157

H0: There is no relationship between all independent variables and FDI inflow in

Malaysia.

H1: There is relationship between all independent variables and FDI inflow

in Malaysia.

H0: β1 = 0 (There is no relationship between market size and FDI inflow in

Malaysia.)

H1: β1 ≠ 0 (There is relationship between market size and FDI inflow in

Malaysia.)

H0: β2 = 0 (There is no relationship between economic growth and FDI inflow in

Malaysia.)

H1: β2 ≠ 0 (There is relationship between economic growth and FDI inflow

in Malaysia.)

H0: β3 = 0 (There is no relationship between quality of infrastructures and FDI

inflow in Malaysia.)

H1: β3 ≠ 0 (There is relationship between quality of infrastructures and FDI


H0: β4 = 0 (There is no relationship between trade openness and FDI inflow in

Malaysia.)

H1: β4 ≠ 0 (There is relationship between trade openness and FDI inflow in

Malaysia.)

H0: β5 = 0 (There is no relationship between exchange rate and FDI inflow in

Malaysia.)

H1: β5 ≠ 0 (There is relationship between exchange rate and FDI inflow in

Malaysia.)

H0: β6 = 0 (There is no relationship between China FDI inflows and FDI inflow in

Malaysia.)

H1: β6 ≠ 0 (There is relationship between China FDI inflows and FDI


H0: β7 = 0 (There is no relationship between inflation rate and FDI inflow in

Malaysia.)

H1: β7 ≠ 0 (There is relationship between inflation rate and FDI inflow in

Malaysia.)


Page 37 of 157

2.5 Conclusion

This research has used 7 independent variables constituting of market size,

economic growth, quality of infrastructure, trade openness, exchange rate, China

FDI inflow, and inflation rate. As supported by previous studies, researchers

assume that those variables are significant in determining the foreign direct

investment inflows in Malaysia. Therefore, researchers will be collecting those

indicators‟ observations from reliable database and plan carefully for the research

methodology so as to obtain a proper analysis to prove what they assumed is

correct and accurate.


Page 38 of 157

CHAPTER 3: METHODOLOGY

3.0 Introduction

Research methodology is a way to systematically solve the research problem. It

may be understood as a science of studying how research is done scientifically. At

each operational step of the research process, researchers are required to choose

from a multiplicity of methods, procedures and models of research methodology

to achieve the objectives. So in this chapter, researchers will list out the various

steps that are generally adopted in researching the determinants of foreign direct

investment along with the logic behind them. What data have been collected and

what particular method has been adopted, why a particular technique of analyzing

data has been used and a host of similar other questions will be answered as

researchers get deeper in the research methodology concerning the research study.

The choices and decisions made in this part of the research take into consideration

the literature review as done in the previous chapter. Literature review tells if

others have used procedures and methods similar to the ones that researcher are

proposing, which procedures and methods have worked well for them, and what

problems they have faced with them. Thus, it allows researcher to better position

in selecting a methodology that is capable of providing a valid answer to our

research questions.

Last but not least, the research design, data collection methods, data processing,

and methods of data analysis will be discussed in the following sub topics.


Page 39 of 157

3.1 Research Design

Researchers use research design as the fundamental directions to carry out the

study. In this study, researchers opted to use quantitative research. Quantitative

research involves a collection of numerical data to answer a specific research

question. In this case, the research is to examine the relationship between the

independent variables which are China FDI inflow (CFDI) as measured by China

FDI (Athukorala & Waglé, 2011; Wang, Wei ,& Liu, 2007 and Salike, 2010),

quality of infrastructure (TL) as measured by telephone line per 100 people

(Asiedu, 2002; Mengistu & Adhikary ,2011), exchange rate (OREER) as

measured by official real effective exchange rate (Hara & Razafimahefa, 2005),

trade openness (TO) as measured by ratio of exports plus imports to GDP

(Demirhan & Masca, 2008; Nurudeen, Wafure & Auta, 2011), inflation rate (INF)

as measured by inflation, consumer price (Demirhan & Masca, 2008), market size

(GDP) as measured by real GDP per US$( Sharma & Bandara, 2010 ; Awan,

Khan , & Zaman, 2010), and economic growth (GDPG) as measured by annual

gross domestic product growth rate (Kahai, 2011 & Ang, 2008) and the

dependent variable which is foreign direct investment inflows in Malaysia.

Besides that, researchers collect data on predetermined instruments to yield

statistical data. It is a more structured data collection technique. Quantitative

research not only provides the summary of the information on the characteristic,

but it is also useful in tracking the trend. There are three type of research design:

exploratory research, descriptive research, and causal research (McDaniel & Gates,

2010). According to McDaniel and Gates (2010), exploratory research is useful

when the research question is unclear to guide the progress of the hypotheses. This

research is also used to develop a better understanding on a problem or

opportunity. Descriptive research explains more on some situation by providing a

measure of the event or activity and it is accomplished by using descriptive

statistic. Meanwhile, causal research is the most complex compare to the other

two. This is because such research design is used to test whether one event or

activity causes another (McDaniel & Gates, 2010). In this research paper,

exploratory research is used for there are some unclear research questions. As


Page 40 of 157

researchers have included some new indicators like China FDI inflow and

infrastructure that have not been used by past researchers on the country of

Malaysia, it was unsure whether or not there will be a significant influence by the

new indicators on Malaysia‟s foreign direct investment inflows.

3.2 Data Collection Methods

In order to study the effects the variables have on the FDI inflows in Malaysia,

researchers had obtained data on the indicators of market size, economic growth,

inflation rate, exchange rate, trade openness, quality of infrastructure and China

FDI inflow from the World Bank database. For the study, data researchers used

the annual times series data from the year 1982 to 2010 which consists of 29

observations. The GDP per US$ is used as the proxy for market size while the

annual GDP growth rate is used as the proxy for economic growth. Besides, the

indicator for inflation is inflation, consumer price while the real effective

exchange rate is the indicator for exchange rate. Meanwhile, researchers employed

the ratio of exports plus imports to GDP as the proxy for trade openness. Quality

of infrastructure, on the other hand, has the indicator of telephone line per 100

people. Last but not least, the FDI inflows in China are used for the China FDI.

In the research project, quantitative research is done since all the data obtained are

quantitative data. Quantitative research as defined by Cohen (1980) (as cited in

Sukamolson, 2005) is the social research that uses empirical methods and

empirical statements which are expressed in numerical terms. Besides,

Sukamolson (2005) defined quantitative research as explaining phenomena by

collecting quantitative data that are analyzed using mathematically based methods.

Since the data researchers obtained from World Bank are in numerical terms, it

qualifies as the quantitative data as indicated in the statement made by

Sukamolson (2005).

Time series data is a collection of observations on the values that a variable takes

over a period such as daily, monthly, quarterly, and annually (Gujarati & Porter,


Page 41 of 157

2009). Researchers used time series data in the study since the data is collected at

a regular time intervals which is from 1982 to 2010. All the data researchers used

are secondary data as the data have already been collected by someone else for

other purposes. The reasons researchers use secondary data is because working

with secondary data is more economic. Since the data had previously been

collected by someone else for other research purposes, there is no need for current

researchers to spend time in collecting the data by themselves. This in turns save

time so that researchers can focus more on the analysis. Besides, it also saves cost

as the data needed can be easily found and obtained from online sites such as

World Bank both at a price or for free.

3.3 Data Processing

In order to make sure accurate data are selected for the analysis, researchers

constantly check through, update and edit the data. Firstly, referral on several past

researches confirms that the indicators researchers choose to use is fully supported

and proved by past researchers. While collecting the data, researchers double

confirm the data and indicators used are in line with what was used by past

researchers previously. As some of the data obtained from World Bank contained

omitted observations (e.g. FDI china only has 29 observations while GDP has 34

observations), researchers had to make some adjustment on it so as to make sure

all observations are in the same observations‟ years. Even when the researchers

key in the data into E-views, the figures are checked on a few times to make sure

that there is no error. This step further improves the accuracy of the data before

the conduction of the analysis.


Page 42 of 157

3.4 Data Analysis

In this research, Electronic Views (Eviews) is used to run and test the regression

analysis.

3.4.1 Eviews

Eviews is one of the most popular econometric packages around. It can be

used for general statistical analysis and econometric analyses, such as

cross-section and panel data analysis, time series estimation and

forecasting. It combines both spreadsheet and relational database

technology with the traditional tasks found in statistical software. It also

can use Windows GUI to combine with a programming language which

displays limited object orientation (Renfro, 2004).

Eviews relies heavily on a proprietary and undocumented file format for

data storage. However, for input and output, it supports numerous formats,

including databank format, Eel formats, PSPP or SPSS, DAP or SAS, Stata,

Rats, and TSP. Furthermore, it can access OECD databases. According to

Startz (2009), Eviews can estimate a regression and show the information

on each estimated coefficient from the Eviews output. In addition to

regression coefficients, Eviews also can provide a great deal of summary

information about each estimated equation.

In the research paper, Eviews is used to run the estimated multiple

regressions model and also to do the diagnostic checking for determining

whether multicollinearity, autocorrelation and heteroscedasticity problems

exist or not. Besides that, researchers also run the model specification test

and normality test using Eviews.


Page 43 of 157

3.4.2 Multiple Linear Regressions

Unlike simple linear regression model, multiple linear regressions model is

linear regression models that contain one dependent variable (Y) and two

or more independent variables (Xi) (Gujarati & Porter, 2009). The

independent variables act as explanatory variables in predicting the result

of dependent variable. An equation of multiple linear regressions models is

as below:

Y = β0 + β1X1i + β2 X2i +…..βkXki+ μi

One of the reasons researchers opted to use multiple linear regressions

model instead of simple linear regression model is most probably because

the outcomes of their estimation does not only involve one independent

variable influencing it. In real life, there are multiple factors influencing

the outcomes. Thus, multiple linear regressions model is used to ensure

that the estimated result does not divert from the actual results. In order to

obtain a rather accurate estimation, 7 variables have been included into the

estimated models by researchers. The fewer variables are omitted from the

estimation, the more accurate results will it be. As follow is the estimated

economic model researchers formed:

FDIt= β0 + β1GDPt + β2GDPGt + β3OREERt + β4TLt + β5TOt + β6INFt +

β7CFDIt

Where FDI refers to foreign direct investment inflows in Malaysia, β is the

coefficients used to explain the degree it will affect FDI. GDP is market

size, GDPG is economic growth, OREER is real effective exchange rate,

TL is infrastructure, TO is trade openness, INF is inflation rate and CFDI

is China FDI inflows.

One of the characteristics of multiple linear regressions function is that the

parameters (βk) in the model should be linear and there is no relationship


Page 44 of 157

among the independent variables. The main reasons researchers need to

avoid any two independent variables with relationship with each other is

because if the two independent variables are highly correlated

(multicollinearity), researchers may be getting biased information from the

model.

In multiple linear regressions models, the β1 and β2 are partial regression

coefficients (with two predictor variables). Partial regression coefficients

indicate how the independent variables (Xi) take effect on dependent

variable (Y), withholding other variables constant. On the other hand, in

examining whether the multiple linear regressions model is fitted with the

data, adjusted R2

is preferred instead of R2. This is because R

2 never

decrease as the number of independent variables included in the model

increases.

3.4.3 F- test Statistic

F- test is a measure on the overall significance of the estimated regression.

It is any statistical test in which the test statistic has an F-distribution under

the null hypothesis. F- test is used when multiples parameters are involved

in the model. It is most commonly used in comparing statistical

models that have been fit to a data set to identify the model that best fits

the population from which the data were sampled.

Statistics F-test helps to analyze data by using the F-test statistic to

determine a P-value that indicates how likely one could have gotten the

results by chance. Therefore, if there is a less than either 1%, 5%, or 10%

chance of getting the observed differences by chance, researchers will

reject the null hypothesis and find it statistically significant for the whole

model. This also means that if the P-value of F-test is lower than 0.01, 0.05,

or 0.1, they will reject the null hypothesis and conclude that it is significant

for the whole model to explain the dependent variable.


Page 45 of 157

3.4.4 T- test Statistic

T-test is probably the most commonly used Statistical Data Analysis

procedure for hypothesis testing of the means of the variables associated

with two independent samples or groups (Lucey, 2002). T-test requires

interval or ratio data and assumes the sample populations have normal

distribution while the variances are equal. This test assess whether the

observed differences between two sample means occurred by chance.

According to Lucey (2002), statistics T-test helps to analyze the data by

using the t-test statistic to determine a P-value that indicates how likely

one could have gotten the results by chance. Therefore, if there is a less

than either 1%, 5%, or 10% chance of getting the observed differences by

chance, researchers will reject the null hypothesis and find a statistically

significant difference between the two groups. This also means that if the

P-value of T-test is lower than 0.01, 0.05, or 0.1, they will reject the null

hypothesis and conclude that there is significance between the independent

variable and dependent variable.

In short, statistics T- test is used to test the significance of each

independent variable to dependent variable.

3.4.5 Diagnostic Checking

As mentioned above, there might be an existence of econometric problems

in the model. Hence, researchers need to conduct several hypotheses

testing to check and detect whether the model is free from multicollinearity,

autocorrelation and heteroscedasticity problems. Furthermore, researchers

will also need to test for model specification and carry out normality test as

well.


Page 46 of 157

3.4.5.1 Model Specification and Normality test

Model specification refers to the determination of which independent

variables should be included in or excluded from a regression equation

(Gujarati & Porter, 2009). Indeed, it can be observed that regression

analysis involve three distinct stages: the specification of a model, the

estimation of the parameters of the model, and the interpretation of these

parameters. Model specification is the first and most critical of these stages.

Researchers‟ estimates on the parameters of a model and their

interpretation of them depend on the correct specification of the model.

Therefore, problems can arise whenever researchers wrongly specify a

model. Model specification error occurs when there involve any omission

of relevant variable or inclusion of unnecessary or irrelevant variable or

due to the wrong functional form (Gujarati & Porter, 2009). When

legitimate variables are omitted from a model, the Ordinary Least Square

estimators of the variable retained in the model will be biased and

inconsistent. That is the variances and standard errors of these coefficients

will be incorrectly estimated and vitiate the actual hypothesis-testing. On

the contrary, the consequences of including irrelevant variables in the

model are less serious. Estimators of the coefficients of the relevant as well

as irrelevant variables will remain unbiased and consistent while the error

variance remains correctly estimated. The only problem is that the

estimated variances tend to be larger than necessary, thereby leading to a

less precise estimation of the parameters. Or in other words, the confidence

intervals tend to be larger than necessary. Test can be used to detect on the

model specification error is Ramsey‟s RESET test.

Right behind the model specification test is another test which also cannot

be ignored. The normality test is used to check whether the error term of

the model is normally distributed or not. If the error term of the model is

not normally distributed, the estimated model will be biased and the

hypothesis testing result will be affected as well. If researchers are dealing

with a small or finite sample size which is less than 100 observations, the

normality assumption assumes a critical role. It not only helps us in


Page 47 of 157

deriving the exact probability distributions of Ordinary Least Square

estimators but also enable us to use the T-test, F-test and other statistical

tests for regression models (Gujarati & Porter, 2009).

Therefore, before proceeding to diagnostic checking and other statistical

tests, researchers have to make sure that their model is well specified and

its error term is normally distributed. In this research paper, Ramsey‟s

RESET test will be used to check on the model specification while Jarque-

Bera normality test will be used to see through the normality of the error

term.

3.4.5.2 Multicollinearity

Researchers have utilized the term independent variable to refer to variable

being used in the forecast or clarification of the value of dependent

variable. This does not mean the independent variables are independent in

a statistical sense. However, there is high probable that most of the

independent variables are correlated.

At the outset, researchers have planned to run some test on the

multicollinearity problem in order to assess whether the independent

variables in the model are highly related among each other or not.

Researchers will be applying covariance analysis to find out whether there

is correlation between the independent variables.

Multicollinearity can lead to a numbers of problems with regression. In

some cases, multicollinearity will cause the regression coefficients to have

a sign opposite that of the actual relationship. Therefore, when there is a

high degree of multicollinearity, researchers cannot rely on the individual

coefficients to interpret the results. This is because it is certain that the

problem had affect the statistical significant of the individual regression

coefficients and the ability to use them to explain the relationships. It is

agreed by Larget (2007) that when multicollinearity is present, important


Page 48 of 157

variables can appear to be non-significant and the standard errors can be

large than it is supposed to be.

To assess multicollinearity, the correlation among the independent

variables should be known. The results are shown both as an individual R-

squared and a Variance Inflation Factor (VIF). When the R-squared and

VIF values are high for any of the variables, the model is affected by

multicollinearity (Motulsky, 2002).

3.4.5.3 Autocorrelation

According to Gujarati and Porter (2009), autocorrelation may be defined

as correlation between members of the observations ordered in time and

place. As in time series data, there might be the correlation between

disturbance terms. Another possible reason that autocorrelation will

occurred is because researchers omitted some important variables from the

model or used the wrong functional form. If autocorrelation occurs in the

stated model, researchers might get bias results.

Therefore, researchers need to check for autocorrelation problem as it

might occur in the studied model if the error is correlated. In detecting the

autocorrelation problem, researchers have selected Breusch-Godfrey Serial

Correlation LM Test to run the test. Since research sample size is small

which has 29 observations, researchers also include informal way

(graphical method) to detect autocorrelation problem. This is because the

sample is small and graphical method will provide a more accurate answer

compare to hypothesis testing, so the final conclusion will be drawn base

on graphical result.


Page 49 of 157

3.4.5.4 Heteroscedasticity

Follow on, researchers run the heteroscedasticity test to test for the

constant variance of error terms. Researchers use ARCH test to identify

the heteroscedasticity problem in this model. As stated by Gujarati and

Porter (2009), model with heteroscedasticity problem have error terms that

do not have a constant variance. There may be larger variance when values

of some independent variables tend to be larger or smaller. Hence, the

model with heteroscedasticity problem will no longer have minimum

variances or be efficient. This will lead to the incorrect conclusion.

As stated in the theory, if heteroscedasticity occurs, it will not be easy to

correct them. If the sample size is large, White‟s Heteroscedasticity-

consistent Variances and Standard Errors can be used to correct the

standard errors of OLS estimators and conduct statistical inference based

on these standard errors (Gujarati & Porter, 2009). Since research sample

size is small which has 29 observations, researchers also include informal

way (graphical method) to detect heteroscedasticity problem. This is

because the sample is small and graphical method will provide a more

accurate answer compare to hypothesis testing, so the final conclusion will

be drawn base on graphical result.

3.5 Conclusion

After identifying both the data and methodology that researcher opted to use, the

analysis on the data will be done in Chapter 4 using the Ordinary Least Square

(OLS) estimator.


Page 50 of 157

CHAPTER 4: DATA ANALYSIS

4.0 Introduction

This chapter revolved the analysis of data that had been collected for the study.

Using multiple linear regression method, researchers analyse the data to identify

which independent variables significantly affect the dependent variable, FDI

inflows to Malaysia (MFDI). On top of that, relationship between the independent

variables and the dependent variable are also figured out. Throughout this chapter,

data analysis would be carried out so as to fulfil both the objectives and

hypothesis which was mentioned in Chapter 1.

4.1 Empirical Result of Multiple Linear Regressions

Model

With the annual data from the year 1982 to 2010, researchers run the model using

E-views and the following results are obtained:

𝑀𝐹𝐷𝐼𝑡 = 2.00E+09- 0.010788GDPt+ 1.53E+08GDPGt-2.32E+09OREERt

se = (2.09E+09) (0.021409) (86462339) (9.89E+08)

p-value = (0.3497) (0.6196) (0.0917)* (0.0289)*

+8122953TLt+3.78E+09 TOt+ 14902615 INFt+ 0.044008CFDIt

se = (1.93E+08) (2.94E+09) (1.26E+08) (0.021564)

p-value = (0.9668) (0.2124) (0.9070) (0.0540)*

n = 29 R2= 0.832961 R 2 = 0.777282 Prob(F-statistic) = 0.000001*

*significant at 0.10 significance level

Following the attainment of the empirical results of multiple linear regressions,

researchers carried out diagnostic checking tests to ensure the error terms of the


Page 51 of 157

multiple linear regressions are normally distributed, the model is correctly

specified and free from multicollinearity, heteroscedasticity and autocorrelation

problems.

4.1.1 Diagnostic Checking of Multiple Linear Regression

Model

Table 4.1.1 Summary of Diagnostic Checking of Multiple Linear

Regressions

Hypothesis Testing p-value

1. Jarque-Bera normality test 0.790615

2. Ramsey‟s RESET test 0.0275

3. Multicollinearity test

3.1 Correlation and Variance Inflation Factor (VIF)

Independent

Variables

Correlation Variance Inflation Factor,

VIF

a. CFDI GDP 0.971965 18.08840342

b. TL TO 0.953759 11.06883186


4. Autoregressive Conditional

Heteroscedasticity (ARCH) test

0.2025

5. Breusch-Godfrey LM test 0.0048

Source: Developed for the research

To find out whether the error terms of the model are normally distributed,

Jarque-Bera normality test was used. Given the p-value of 0.790615 is

more than α, 0.10, researchers conclude that the error terms of this multiple

linear regressions model are normally distributed.


Page 52 of 157

Next, Ramsey‟s RESET test was performed to determine whether the

model is correctly specified. The researchers easily verify that the p-value

(0.0275) was less than the significance level of 0.10, indicating that the

multiple linear regressions model is mis-specified. Researchers, however,

did come to a conclusion that Ramsey‟s RESET test does not apply in this

study as the sample size was too small with only 29 observations. A test

such as RESET will only provide an accurate conclusion on the model

specification if the sample size is reasonably large.

Follow on, researchers proceed with multicollinearity testing. Base on the

precedent studies, the official real exchange rate, OREER is expected to

have relationship with the trade openness, TO. Adding on to the high R-

squared value, only one independent variable was found to be significant.

Hence, this further increases the researchers‟ suspicion that the model has

multicollinearity problem. Researchers then carried out correlation test and

variance inflation factor (VIF) to see which variables are highly correlated.

The correlation table shows that official real exchange rate and trade

openness has a low correlation of 0.828626 and a variance inflation factor

of3.191024765 which put them in a no serious multicollinearity problem

range of 1 and 5. However, the pair-wide correlations of gross domestic

product (GDP) and China foreign direct investment inflows (CFDI); and

telephone lines (per 100 people) (TL) and trade openness (TO) shows high

correlations that is 0.971965 and 0.953759 respectively. The variance

inflation factor, VIF calculated also indicated these 2 pairs have high value

of variance inflation factor which are 18.08840342 and 11.06883186

respectively. Since the values are more than 10, the pairs of gross domestic

product (GDP) and China foreign direct investment inflows (CFDI); and

telephone lines (per 100 people) (TL) and trade openness (TO) are

concluded as having serious multicollinearity problem.

Despite Jarque-Bera normality test proved that the error terms are

normally distributed, it does not free the model from heteroscedasticity and

autocorrelation problems. In order to detect heteroscedasticity and


Page 53 of 157

autocorrelation problems, hypothesis checking has to be carried out.

Autoregressive Conditional Heteroscedasticity (ARCH) test is chosen to

detect heteroscedasticity problem while Breusch-Godfrey LM test is used

to detect autocorrelation problem. Researchers use a maximum of 6 lag

lengths to detect heteroscedasticity and autocorrelation problems as the

sample size is small. If researchers were to include too many lag length

inside, the degree of freedom would decrease and incurred inappropriate

hypothesis testing result. Among the 6 lag length, researchers chose p-

value from the lag length with the lowest Akaike Information Criterion

(AIC) to make decision. Owing to the small sample size, graphical method

is deemed to be more appropriate than hypothesis testing in detecting both

the problems. Somehow, we do both hypothesis testing and graphical

method to compare and detect the problems.

Autoregressive Conditional Heteroscedasticity (ARCH) test indicates that

the model is free from heteroscedasticity problem as the p-value is 0.2025,

greater than the critical value of 0.10. From this, researchers conclude that

the model does not have heteroscedasticity problem. But somehow, the

model does appear to have autocorrelation problem. The p-value of the

model which is 0.0048 is smaller than the significance level of 0.10. Using

the hypothesis testing method, researchers conclude that the model has

autocorrelation problem but not heteroscedasticity problem.

However, the residual graph shows that the model has both

heteroscedasticity and autocorrelation problem. The high volatility of error

terms point out the existence of heteroscedasticity problem. In addition,

the decreasing trend of the error terms indicates the autocorrelation

problem.

In conclusion, the error terms of multiple linear regressions model are

normally distributed, but the model specification is incorrect and has

multicollinearity, heteroscedasticity and autocorrelation problems.

Researchers conclude that the estimated coefficient values are biased and

inefficient for the problems encountered. The standard errors, t-statistics


Page 54 of 157

values, F-statistic value and p-value of individual independent variables

and whole estimates are inaccurate.

4.2 Problem Solving

4.2.1 Problem Solving of Model Specification

Subsequently, researchers tried to solve the model specification problem

by changing the form of dependent variable data. Log is added onto the

dependent variable and a Log-Lin model is formed as described in the

following:

lnMFDIt = β8 + β9GDPt + β10GDPGt + β11OREERt + β12TLt + β13TOt +

β14INFt + β15CFDIt

The empirical results of the Log-Lin model are

𝑙𝑛 𝑀𝐹𝐷𝐼𝑡 = 19.37893 + 2.91E-12GDPt + 0.054179GDPGt

se = (0.814918) (8.36E-12) (0.033766)

p- value = (0.0000)* (0.7315) (0.1235)

- 0.803537OREERt - 0.056388TLt + 2.654155TOt

se = (0.386136) (0.075309) (1.148582)

p- value = (0.0499)* (0.4623) (0.0311) *

+ 0.066051INFt + 4.89E-12 CFDIt

se = (0.049206) (8.42E-12)

p- value = (0.1938) (0.5675)

n = 29 R2

= 0.813519 R 2 = 0.751359 Prob(F-statistic) = 0.000002*

* significant at 0.10 significance level


Page 55 of 157

4.2.1.1 Diagnostic Checking of Semi-Logarithmic: Log-Lin Model

Table 4.2.1.1: Summary of Diagnostic Checking of Semi-logarithmic:

Log-Lin Model





The Jarque-Bera normality test shows that the model‟s error terms are

normally distributed for the p-value (0.301500) is more than α, 0.10. Other

than that, Ramsey‟s RESET test indicates that the model specification is

correct with its p-value (0.5140) more than the significance level of 0.10.

In short, researchers were able to solve the model specification problem by

adding on log to the dependent variables.

4.2.1.2 Problem Solving of Multicollinearity

Although the model specification problem is solved, multicollinearity

problem remains because researchers only changed the form of the

dependent variable data and not the form of independent variables data.

Multicollinearity problem may also be easily solved by adding log to the

independent variables. However, it is not possible due to the negative

figures in the independent variable gross domestic product growth. Instead,

researchers chose to split the single model into two models. Given gross

domestic product (GDP) and China foreign direct investment inflows

(CFDI); and telephone line (per 100 people)(TL) and trade openness (TO)

have serious multicollinearity relationships, we opted for trial and error

method to pick out one of the variables from the pair variables and form

four new models.


Page 56 of 157

Model 1: lnMFDIt = β16 + β17GDPGt + β18OREERt + β19TOt + β20INFt +

β21CFDIt

Model 2: lnMFDIt = β22 + β23GDPGt + β24OREERt + β25TLt + β26INFt +

β27CFDIt

Model 3: lnMFDI t= β28 + β29GDPt + β30GDPGt + β31OREERt + β32TOt +

β33INFt

Model 4: lnMFDIt = β34 + β35GDPt + β36GDPGt + β37OREERt + β38TLt +

β39INFt

Follow on, researchers choose the best model out of the four models

formed by first comparing the significance of the whole model, then their

adjusted R2, and lastly, the number of independent variables significant in

the model.

Table 4.2.2: Summary of Comparisons among 4 Models

Model 1 Model 2 Model 3 Model 4

p-value of

independent

variables

GDP 0.0009* 0.0028*

GDPG 0.0177* 0.0005* 0.0268* 0.0009*

OREER 0.0571* 0.5819 0.0726* 0.7482

TL 0.0330*

TO 0.0005* 0.0085* 0.0013*

INF 0.0986* 0.0404* 0.1152 0.0539*

CFDI 0.0004* 0.0008*

Adjusted R2 0.766708 0.707391 0.749916 0.676651

Prob(F-statistic) 0.000000* 0.000002* 0.000000* 0.000005*




Page 57 of 157

From the result, all models are found to be significant in explaining

Malaysia FDI inflows. Thus, researchers chose the best model based on the

highest adjusted R2. Model 1 out of all the other four models proved to be

more accurate in explaining the data. On top of that, all independent

variables are found significant in the model. Now that the multicollinearity

problems have been solved, researchers proceed onto diagnostic checking

of model 1.

4.3 Empirical result of Model 1

Model 1: ln MFDIt = β16 + β17GDPGt + β18OREERt + β19TOt + β20INFt +

β21CFDIt

Model 1 is chosen which includes economic growth, official real exchange

rate, trade openness, inflation rate and China FDI inflows as independent variables

Malaysia FDI inflows as dependent variable.

ln 𝑀𝐹𝐷𝐼𝑡 = 19.62773 + 0.068305GDPGt - 0.693089OREERt + 1.860519TOt +

se = (0.705766) (0.026728) (0.346083) (0.461854)

p-value = (0.0000)* (0.0177)* (0.0571)* (0.0005)*

0.077766INFt + 7.66E-12 CFDIt

se = (0.045178) (1.84E-12)

p-value = (0.0986)* (0.0004)*

n = 29 R2


* significant at 0.10 significance level


Page 58 of 157

4.3.1 Diagnostic checking of Model 1

Table 4.3: Summary of Diagnostic Checking of Model 1






0.6510



Jarque-Bera normality test shows that the error terms of Model 1 are

normally distributed as the p-value (0.274259) is greater than the

significance level of 0.10. Ramsey‟s RESET test also indicates that the

model is correctly specified since the p-value of 0.7028 is greater than the

significance level of 0.10. The model is cleared from multicollinearity

problems after it is split. Moreover, ARCH test shows that the model no

longer has heteroscedasticity problem as the p-value of 0.6510 is greater

than the 0.10 significant level. However, the Breusch-Godfrey LM test

indicates that the model has autocorrelation problem to which the error

terms follow autoregressive order of 5, AR (5) and the p-value (0.0210) is

smaller than the 0.10 significant level. To conclude, ARCH and Breusch-

Godfrey LM hypothesis testing proved the model has autocorrelation

problem but not heteroscedasticity problem. Quite the reverse, the residual

graph provides a contrary result. From the graph, it is obvious that the

volatility of the error terms is high and is showing descending trend. Thus,

researchers conclude that the model has heteroscedasticity and

autocorrelation problems.


Page 59 of 157

In short, the error terms of Model 1 are normally distributed and the model

is correctly specified. Furthermore, it has no multicollinearity problem

only heteroscedasticity and autocorrelation problems.

4.3.2 Heteroscedasticity Problem Solving of Model 1

Based on the residual graph of model 1, researchers identified the model

with heteroscedasticity problem but were not aware of the severity of the

problem as Breusch-Godfrey LM hypothesis testing indicated that the

model has no heteroscedasticity problem. There was no empirical result

allowing researchers to know the degree of heteroscedasticity problem in

the model. In face of this problem, the estimated coefficient values are

assumed to have become inefficient. So in order to obtain a better result,

researchers used White‟s Heteroscedasticity-consistent Variances and

Standard Errors to minimize the heteroscedasticity problem. By using this

method, not only the standard error of the dependent variables, t-statistic as

well as p-value of the independent variables will also be largely improved.

𝑙𝑛 𝑀𝐹𝐷𝐼𝑡 = 19.62773 + 0.068305GDPGt - 0.693089OREERt

se = (0.522269) (0.035372) (0.304931)

p-value = (0.0000)* (0.0659)* (0.0327)*

+ 1.860519TOt + 0.077766INFt + 7.66E-12 CFDIt

se = (0.483130) (0.035864) (1.69E-12)

p-value = (0.0008)* (0.0407)* (0.0001)*

n = 29 R2




Page 60 of 157

4.3.3 Autocorrelation Problem Solving of Model 1

Researchers were unable to solve the autocorrelation problem in model 1

due to the limited knowledge. The autocorrelation problem in model 1 is

considered serious in which the error terms follow autoregressive order 5,

AR (5). In conclusion, model 1 still has heteroscedasticity and

autocorrelation problems.

4.4 Empirical result of Model 5

Model 5: lnMFDIt= β40 + β41GDPt + β42TLt

The above is model 5, another model formed using the independent variables

picked out from the original multiple linear regressions model in order to solve

multicollinearity problem.

𝑙𝑛 𝑀𝐹𝐷𝐼𝑡 = 20.37068 + 6.67E-12GDPt + 0.049327TLt

se = (0.351983) (2.87E-12) (0.031772)

p-value = (0.0000)* (0.0283)* (0.1326)

n = 29 R2



4.4.1 Diagnostic Checking of Model 5

Table 4.4: Summary of Diagnostic Checking of Model 5





Page 61 of 157



0.6080



Based on Jarque-Bera normality test, the error terms of Model 5 are found

to be normally distributed because the p-value of 0.110847 is greater than

the 0.10 significant level. On the contrary, Ramsey‟s RESET test shows

that the model is mis-specified as the p-value (0.0175) is smaller than the

significance level of 0.10.Follow on, in detecting the heteroscedasticity

problem, the ARCH hypothesis testing shows that model 5 has no

heteroscedasticity problem. The p-value 0.6080 is larger than the 0.10

significant level. Meanwhile, Breusch-Godfrey LM test indicates that the

model has autocorrelation problem for the p-value 0.0292 is less than the

significance level of 0.10. However, the residual graph shows that the error

terms have high volatility and overall shows that there is trend in the series

of error terms. Researchers then conclude that model 5 has both

heteroscedasticity and autocorrelation problems.

In short, model 5 is incorrectly specified and has autocorrelation and

heteroscedasticity problems. These serious problems are found to be

partially due to the omission of important variables in model 5.

4.4.2 Problem Solving of Model 5

Despite the severeness of the heteroscedasticity problem in model 5,

researchers tried to minimize it using White‟s Heteroscedasticity-

consistent Variances and Standard Errors. It is impossible to solve

heteroscedasticity problem in this case. Given that graphical method is an

informal way to detect the heteroscedasticity problem, it cannot provide

empirical result about how severe the heteroscedasticity problem is.


Page 62 of 157

Hence, researchers can only use White‟s Heteroscedasticity-consistent

Variances and Standard Errors to minimize the problem. Below is the

result after the adjustment using White‟s Heteroscedasticity-consistent

Variances and Standard Errors method:


se = (0.328888) (2.75E-12) (0.031196)

p-value = (0.0000)* (0.0225)* (0.1259)

n = 29 R2



In spite of the improved result, the model still faces heteroscedasticity

problem. Restraint by the researchers‟ limited knowledge, model mis-

specification, heteroscedasticity and autocorrelation problems remains

unsolved for the sample size is too small and important independent

variables were omitted.

4.5 Interpretation of Multiple Linear Regression Results

of Model 1 and 5

Model 1 and 5 is undeniable a better version of the original model as there is no

multicollinearity problem and the heteroscedasticity problem is minimized.

Therefore, it can be concluded that the result of model 1 and 5 is more reliable

than the result of the original model.


Page 63 of 157

4.5.1 Interpretation of Model 1 Result

𝑙𝑛 𝑀𝐹𝐷𝐼𝑡 = 19.62773 + 0.068305GDPGt- 0.693089OREERt +

se = (0.522269) (0.035372) (0.304931)

p-value = (0.0000)* (0.0659)* (0.0327)*

1.860519TOt + 0.077766INFt + 7.66E-12 CFDI

se = (0.483130) (0.035864) (1.69E-12)

p-value = (0.0008)* (0.0407)* (0.0001)*

n = 29 R2



Based on the E-view result, independent variables like gross domestic

product growth (GDPG), official real exchange rate (OREER) trade

openness (TO) inflation rate (INF) and China FDI inflows (CFDI) are

found to be significant as their p-values is less than the 0.10 significant

level. Overall, the model is significant as the p-value in whole which is

0.000000 is less than the 0.10 significance level. The adjusted R2which is

quite high, indicates that 76.6708% of the variation in Malaysia‟s Foreign

Direct Investment inflows, MFDI can be explained by the total variation in

gross domestic product growth (GDP), official real exchange rate

(OREER), trade openness (TO), inflation rate (INF), and China FDI

inflows (CFDI) taking into account the sample size and the number of

independent variables. Table 4.5.1 shows the interpretation of the

significant independent variables estimated coefficient values.

Table 4.5.1: Interpretation of the Significant βs of Model 1.

βs Interpretation

β 16= 19.62773 The value of 19.62773 is the intercept of the line, indicating

the average level of Malaysia FDI inflows is 1962.773%

when the level of gross domestic product growth (GDPG)

official real exchange rate(OREER), trade openness(TO)

inflation rate(INF) and China FDI inflows(CFDI) are zero.


Page 64 of 157

However, this intercept value is not meaningful and can be

ignored.

β 17 = 0.068305 If gross domestic product growth (GDPG) is predicted to

increase by 1%, Malaysia FDI inflows (MFDI) will increase

by 6.8305%, holding the value other variables constant.

β 18 = - 0.693089 If official real exchange rate (OREER) are predicted to

increase by RM1 per US$ (Malaysia Ringgit depreciate),

Malaysia FDI inflows (MFDI) will decrease by 69.3089%,

holding the value other variables constant.

β 19 = 1.860519 If trade openness (TO) are predicted to increase by 1 unit,

Malaysia FDI inflows (MFDI) will increase by 186.0519%,


β 20 = 0.077766 If inflation rate (INF) are predicted to increase by 1 %,

Malaysia FDI inflows (MFDI) will increase by 7.7766%,


β 21 = 7.66E-12 If China FDI inflows (CFDI) are predicted to increase by

US$ 1, Malaysia FDI inflows (MFDI) will increase by 7.66E-

10 %, holding the value other variables constant.


4.5.2 Interpretation of Model 5 Result


se = (0.328888) (2.75E-12) (0.031196)

p-value = (0.0000)* (0.0225)* (0.1259)

n = 29 R2



The E-view result shows that the independent variables telephone line (per

100 people) (TL) is significant because its p-value (0.1259) is more than

the 0.10 significance level. Another independent variable, gross domestic

product (GDP) is significant as its p-values, 0.0225 is less than the level of


Page 65 of 157

significance of 0.10. Overall, the model is significant as the overall p-value

which is 0.000448 is less than the 0.10 significance level. The adjusted

R2which is low, means that 40.4864% of the variation in Malaysia‟s

Foreign Direct Investment inflows (MFDI) can be explained by the total

variation in gross domestic product (GDP) and telephone line (per 100

people) (TL) taking into account the sample size and the number of

independent variables. Table 4.5.2 shows the interpretation of the

significant independent variables estimated coefficient values.

Table 4.5.2: Interpretation of the Significant βs of Model 5.

βs Interpretation

β 40 = 20.37068 The value of 20.37068 is the intercept of the line. It indicates

the average level of Malaysia FDI inflows (MFDI) is

2037.068% when the level of gross domestic product (GDP)

and telephone line (per 100 people) (TL) are zero. However,

this intercept value is not meaningful and can be ignored.

β 41= 6.67E-12 If gross domestic product (GDP) is predicted to increase by

US$1, Malaysia FDI inflows (MFDI) will increase by 6.67E-

10%, holding the value other variables constant.


4.6 Conclusion

Researchers split the original multiple regression model into two models

constituting Model 1 and Model 5 in order to solve the multicollinearity problem.

Model 1 has serious heteroscedasticity and autocorrelation problems whereas

Model 5 has model mis-specification, heteroscedasticity and autocorrelation

problem. Researchers reduce heteroscedasticity problem by using White‟s

Heteroscedasticity-consistent Variances and Standard Errors in order to obtain a

better result. However, they were unable to solve the autocorrelation problem in

both Model 1 and 5 due to the small sample size and limited knowledge. In

addition, researchers also couldn‟t solve the model mis-specification problems in


Page 66 of 157

Model 5 as the model omitted some important independent variables. Due to all

these limitations, researchers were unable to get the best result.

Researchers then chose to interpret the result of Model 1 and 5. It is found that

telephone line (per 100 people) (TL) is insignificantly affecting Malaysia FDI

inflows (MFDI) at 0.10 significant level. This means that this variable has no

relationship with Malaysia FDI inflows (MFDI). On the other hands, it has also

come to the researchers‟ realization that gross domestic product growth (GDPG),

official real exchange rate (OREER), trade openness (TO), inflation rate (INF),

China FDI inflow (CFDI) and gross domestic product (GDP) significantly affect

Malaysia FDI inflows (MFDI) and have relationship with Malaysia FDI inflows

(MFDI) at the significance level of 0.10. Official real exchange rate (OREER) has

negative relationship with Malaysia FDI inflows (MFDI) while gross domestic

product growth (GDPG), trade openness (TO), inflation rate (INF), China FDI

inflows (CFDI) and gross domestic products (GDP) have positive relationship

with Malaysia FDI inflows (MFDI).


Page 67 of 157

CHAPTER 5: DISCUSSION, CONCLUSION AND

IMPLICATIONS

5.0 Introduction

In this tough economic time, foreign direct investment definitely plays an

important role in Malaysia economic growth. As investigated, many factors are

found to affect investors‟ decision on the country to invest in and how policy

makers attract more FDI inflow to Malaysia. In the previous chapter, researchers

run diagnostic checking tests, F-test, and T-test to examine the significance of the

independent variables on the dependent variable - FDI of Malaysia. In this chapter,

researchers will be comparing the major findings with the past research papers and

also state the implications of this study to both policy makers and practitioners.

Last but not least, this chapter will also comprise the limitations of study and

recommendations for future researches.

5.1 Summary of Statistical Analyses

In chapter 4, researchers started out with the F-test to see whether the multiple

linear regressions model is significant. After proving the model to be significant,

diagnostic checking was performed, including Jarque-Bera normality test to see

whether the error terms of the model are normally distributed; Ramsey‟s RESET

test to check whether the model is correctly specified; Multicollinearity test to find

out whether there is multicollinearity problem between the variables;

Autoregressive Conditional Heteroscedasticity (ARCH) test to investigate whether

heteroscedasticity problem exist or not; and lastly, Breusch-Godfrey LM test to

find out whether autocorrelation problem do exist in this model.


Page 68 of 157

Based on the statistical results, it is proven that the model specification of this

multiple linear regressions model was incorrect. In order to solve this problem,

researchers changed the form of the dependent variable data by adding on log,

turning it into a semi-logarithmic or so called the Log-Lin model. Also, results

show this model consist of heteroscedasticity, multicollinearity, and

autocorrelation problems. In face with the multicollinearity problem, researchers

split the original multiple linear regressions model into two new models, Model 1

and Model 5. As seen from the results, both of these models are significant but

Model 1 has serious heteroscedasticity and autocorrelation problems, while Model

5 has model specification, heteroscedasticity and autocorrelation problems.

Researchers used White‟s Heteroscedasticity-consistent Variances and Standard

Errors to obtain better result by reducing the heteroscedasticity problem. However,

autocorrelation problem in both Model 1 and Model 5 and model specification

problem in Model 5 was still unable to be solved due to the limited knowledge.

In a nutshell, researchers have chosen Model 1 and Model 5 for interpretation.

Results show that quality of infrastructure (telephone line per 100 people) is

insignificant to Malaysia FDI inflows at 0.10 significance level, while other

independent variables such as economic growth (gross domestic product growth),

exchange rate (official real exchange rate), trade openness (trade openness),

inflation rate (consumer prices), China FDI inflows (China FDI inflows) and

market size (gross domestic product) are significant to Malaysia FDI inflows at

0.10 significance level. The details will be discussed as follows.

Table 5.1: Decision for the Hypotheses of the Study

Hypotheses of the study Decision

i. H0: There is no relationship between all

independent variables and FDI inflow in

Malaysia.

H1: At least one independent variable has

relationship with FDI inflow in Malaysia.

Reject H0


Page 69 of 157

ii. H0: There is no relationship between economic

growth and FDI inflow in Malaysia.

H1: There is relationship between economic

growth and FDI inflow in Malaysia.

Reject H0

iii. H0: There is no relationship between market size

and FDI inflow in Malaysia.

H1: There is relationship between market size


Reject H0

iv. H0: There is no relationship between China FDI

inflows and FDI inflow in Malaysia.

H1: There is relationship between China FDI


Reject H0

v. H0: There is no relationship between exchange

rate and FDI inflow in Malaysia.

H1: There is relationship between exchange rate


Reject H0

vi. H0: There is no relationship between inflation

rate and FDI inflow in Malaysia.

H1: There is relationship between inflation rate


Reject H0

vii. H0: There is no relationship between quality of

infrastructures and FDI inflow in Malaysia.

H1: There is relationship between quality of

infrastructures and FDI inflow in Malaysia.

Do not reject H0

viii. H0: There is no relationship between trade

openness and FDI inflow in Malaysia.

H1: There is relationship between trade openness


Reject H0



Page 70 of 157

5.2 Discussions of Major Findings

5.2.1 Market Size

Market size measured by gross domestic product plays an important role in

this study because it indicates how well a country‟s population demand for

the output. It is important for foreign investor to determine whether to

invest or not from the view of market opportunity.

The hypothesis testing of market size in this research paper shows that

gross domestic product is significant and positively affects Malaysia

foreign direct investment inflows at the significance level of 0.10. This

result is consistent with previous researchers like Quer and Claver (2007),

Rodriguez and Pallas (2008), Vadlamannati, Tamazian and Irala (2009)

and Trevino and Mixon Jr. (2004) which also uses gross domestic product

as the indicator for market size.

This proves that Malaysia FDI inflows will increase given the level of

gross domestic product increase. This is parallel to the study of Sharma &

Bandara (2010) who stated that larger market size will attract investors

with ease. Charkrabarti (2001) (as cited in Moosa & Cardak, 2006) also

supported the findings as it explains large market size means the resources

of the country will be utilized more efficiently and exploitation of

economies of scale. Hence, it is the key for investors who aim for long

term investment.

In contrast, Dermihan and Masca (2008) argued that gross domestic

product is not suitable to be an indicator for market size. Both the

researchers suggested that GDP per capita or GDP growth will be more

appropriate as the indicator for market size. Since this argument is not

backed by many research, gross domestic product is consider preferable as

the indicator for market size in this study.


Page 71 of 157

5.2.2 Economic Growth

In the research paper of Benacek, Gronicki, Holland, and Sass (2000), it is

mentioned that economic growth itself has been identified frequently as an

important determinant of FDI inflow into the host countries. In the

researchers‟ case, it was similar as the test results in chapter 4 proved that

economic growth was statistically positive significant to foreign direct

investment inflows to Malaysia.

The result is consistent with the past researches done by Dunning (1995),

Zhang (2001), Chakraboty and Basu (2002), Moosa (2002), and Hansen

and Rand (2006). In their research papers, they argues that MNCs with

certain ownership advantages will invest in another country with locational

advantages, and both advantages can be captured effectively by

“internalizing” production through FDI. Thus, high GDP growth rate

which represents soundness and stability of economic policies and the

effectiveness of the government institutions will definitely attract foreign

direct investment as it allows them to locate new profit opportunities.

On the contrary, researchers also found the results to be against the study

done by Kahai (2011). In Kahai‟s (2011) study, data from 1998 and 2000

for fifty-five developing countries were employed to examine both the

traditional and non-traditional determinants of foreign direct investment

flowing to developing countries. A significant relationship between

economic growth (as measured by the annual real GDP growth rate) and

FDI was failed to be established in that particular study.

5.2.3 Exchange Rate

Researchers found that the official real exchange rate significantly affects

Malaysia FDI inflows and has a negative relationship with Malaysia FDI

inflows at the 0.10 significance level. This result is in line with the study


Page 72 of 157

done by Aqeel and Nishat (2005). In that study, it is verified that foreign

direct investment increase as exchange rate appreciates in the host country.

Conversely, when the exchange rate of the country depreciates, the FDI of

that particular country decreases as well. Furthermore, Campa (1993)

backed up by stating it is because the investors believe that an appreciation

of the exchange rate will likely increase the future profitability in terms of

the home currency.

On the other hand, multiple researchers found the opposite results. In the

study by Froot and Stein (1991), Klein and Rosengren (1994), Guo and

Trivedi (2002) and Kiyota and Urata (2004), depreciation of exchange rate

in the host country led to the increase in FDI inflows of the host country.

This is because the high currency value reduces the capital of the

investment.

This is inconsistent with the result of this study which displays that when

exchange rate appreciates in the host country, FDI inflows of the host

country will follow along and increase. The reason is because the investors

have high expectation on the economy and the returns.

5.2.4 Quality of Infrastructure

In this research study, the independent variable - quality of infrastructure -

stood out as it is found to be insignificant at the significance level of 0.10.

This means that there is no significant relationship between quality of

infrastructure and foreign direct investment inflows to Malaysia. Or in

other words, quality of infrastructure is not a determinant of Malaysia FDI

inflows.

Infrastructure covers many dimensions, ranging from physical assets such

as roads, sea ports, railways, and telecommunications, to institutional

development, such as accounting and legal services. As mentioned by


Page 73 of 157

Asiedu (2002), a good measure of infrastructure development should take

into account both the availability and reliability of infrastructure. However,

in this case, the measure employed by researchers falls short since it

captures only the availability aspect of infrastructure. This is because

quantitative data on the reliability of infrastructure (such as the frequency

of telephone or power outage) are not very readily available for most

developing countries.

This view is further supported by Quazi (2007), who could not established

positive and significant relationship between infrastructure (measured as

the number of telephones per 1,000 people) and FDI in his study. The

authors admitted that „it is plausible that their proxy variables - the natural

log of the number of telephones available per 1,000 people and the adult

literacy rates, respectively, perhaps inadequately capture their true effects

on FDI. Moreover, Addison et al. (2006) as cited in Rehman, Ilyas, Alam,

and Akram (2011) acknowledge such promotional impact only for

developed nations but, on the other hand, such situation does not exist for

developing countries.

5.2.5 Trade Openness

Based on the results in Chapter 4, trade openness is found to have a

significant positive relationship with Malaysia FDI inflows at the 0.10

significance level. This result is on par with the study done by Moosa and

Cardak (2006), Demirhan and Masca (2008), and Sawkut, Boopen, Taruna,

and Vinesh (2009) whom verified trade openness is significant and have a

positive effect on the inflows of FDI. A country‟s willingness to accept

foreign direct investment is important to the FDI of that particular country.

Chantasasawat, Fung, Iizaka, and Siu (2010), Awan, Khan, and Zaman

(2011) and Srinivasan (2011) further support the results with the statement

that a country with a higher degree of trade openness will lead to an

increase in the foreign direct investment. This shows that investors are


Page 74 of 157

more likely to make investment in those countries which have opened up

to the outside world.

On the other hand, the result obtained from this paper differs from some

researchers like Busse and Hefeker (2007) and Kolstad and Villanger

(2008). In their studies, trade openness is insignificant and negatively

affects FDI inflows. Study‟s result from Goodspeed, Martinez-Vazquez,

and Zhang (2006) also concludes that the effect of trade openness on FDI

is inconclusive.

5.2.6 Inflation rate

This research concluded that inflation rate is a significant determinant of

Malaysia FDI inflows. Based on the result, there is a positive relationship

between inflation rate and Malaysia FDI inflows as supported by past

researchers like Srinivasan (2011) and Addison and Heshmati (2003). In

the past research paper, Addison and Heshmati (2003) mentioned that

higher inflation rate indicates higher price levels which may lead to the

increased production activities of the host country. This then will attract

more foreign firms to invest in the host country for the increased expected

level of profitability. In addition, as stated by Srinivasan (2011), higher

inflation may lead to an increase in product price, which in turns decrease

the demand for host country‟s money. As currency of host country

depreciates, it attracts larger FDI inflows into Malaysia since its capacity

to invest is increased through the reduced cost of capital.

On the other hand, the results are found to be inconsistent with the some of

the past researches done. Despite researchers like Shamsuddin (1994),

Asiedu (2002), Demirhan and Masca (2008), and Azam (2010) have

proven inflation rate to be statistically significant to the FDI inflows, it is

also found that there is a negative relationship between inflation rate and

foreign direct investment. Asiedu (2002) mentioned that a low inflation


Page 75 of 157

rate is taken as a sign of internal economic stability in the host country,

reflecting a lesser degree of uncertainty which encourage foreign direct

investment.

5.2.7 China FDI inflows

According to the statistical results, it shows that China FDI inflows

significantly affect Malaysia FDI inflows. In fact, there is a positive

relationship between China FDI inflows and Malaysia FDI inflows.

This result is different as compared to some past researches. For instance,

Salike (2010) found that there is a high degree of crowding out effect by

China FDI inflows on some other countries‟ FDI which mean that there is

a negative relationship between level of China FDI inflows and level of

FDI inflows to other countries. Other than that, Wang, Wei and Liu (2007)

found that there are significant positive and negative effects of China FDI

inflows based on different countries.

Despite all those, the results of this study are also strongly supported and is

consistent with some past researchers such as Athukorala and Waglé

(2011), Chantasasawat, Fung, Iizaka and Siu (2004), Eichengreen and

Tong (2007), and Eichengreen and Tong (2006). These researches showed

that an increase in the FDI inflows to China would raise the level of FDI

inflows to the other countries. This is due to the production linkages

between China and other countries such as Malaysia taking place in the

form of further fragmentation and specialization of the production process.

Therefore, when components and parts are traded between China and

Malaysia, an increase in China‟s foreign direct investment will positively

relates to an increase in the other countries‟ foreign direct investment. In

addition, as China‟s economy grows, its market will increase and the need

for minerals and resources will rise too. As a result, multinational

enterprises from other countries choose to invest in Malaysia in order to


Page 76 of 157

extract minerals and resources, and then export it to China which is in need

of the whole spectrum of raw materials.

5.3 Implications of the Study

Practically, this research paper provides an insight on decision making for the

investors, policy makers, and practitioners such as Federal Government, Bank

Negara Malaysia. It plays an important role in determining the ways to attract

more foreign direct investment inflow to Malaysia.

Researchers found that gross domestic product growth, official real exchange rate,

trade openness, inflation rate, China foreign direct investment inflow and gross

domestic product significantly affect Malaysia FDI inflows. From the investors

perspectives, increase or decrease in gross domestic product growth can predict

the future development of the country. It also tells investors whether the country is

worth to invest in for long term given that the country‟s development is trending

well.

Moreover, research indicates the government on setting trade barrier in order to

increase trade openness level in Malaysia to attract more trade into the country.

According to Gao (2005), increase in trade openness level will probably lead to

the shifting of advanced technology for infrastructure and communication from

developed country to developing country. This will help in refining the

infrastructure and communication technology of Malaysia, and hence, making

import or export in the country more efficient. Country‟s economic growth will

also increase due to the ability of outsourcing and expansion of products to other

countries.

Besides that, gross domestic product also plays an important role in determining

the foreign direct investment inflow to Malaysia. Gross domestic product

determines the ability of a country‟s population in the demand of outputs.

Investors such as foreign firms or companies will most likely take into


Page 77 of 157

consideration of this factor since sales is based on the country‟s demand. Thus,

this portrays as an indication for the government to refine the economic policy,

increasing the country‟s population income so as to increase the gross domestic

product of the country.

Inflation rate which turns out to be positively significant to the FDI inflows in

Malaysia did not came as a surprise as well. Srinivasan (2011) stated that higher

inflation rate means higher price level of product. And this would attract investors

who aim for higher profit with the thoughts of higher expected level of profit

return. Therefore, Bank Negara Malaysia and the government need to plan

appropriately for the fiscal or monetary policy so as to control the inflation rate

that might attract investors‟ behavior.

The study also discovered that China foreign direct investment inflows are

positively related to Malaysia FDI inflows. According to Chantasasawat, Fung,

Iizaka, and Siu (2004), China‟s low cost of resources is the reason many

foreigners invest and demand for it. And this drastically increasing demand of

resources has led to China‟s needs for more mineral to fulfill the market need.

Malaysia which is one of the developing countries with mineral resources then

becomes the place for foreign investors to invest and bring mineral to China. This

means governments of Malaysia need to discover more low cost resources so as

improve the quality of infrastructure to attract potential investors. Besides,

Malaysia also needs to keep good diplomatic relation with China as it is beneficial.

Lastly, the study showed official real exchange rate is significant but negatively

related to Malaysia FDI inflows. When the exchange rate of Malaysia per

US$ dollar increases, the amount of money foreigners need to use in exchange for

a certain amount of Ringgit Malaysia is lesser. Therefore, investment in Malaysia

will become more worthy as higher profit can be earn when the exchange rate

becomes higher. It is important for Bank Negara Malaysia to set up an ideal

official real exchange rate to attract more FDI inflow to Malaysia.


Page 78 of 157

5.4 Limitations of the Study

The very first limitation of the study is the sample size of the research which is too

small with only 29 years. Annual data from year 1982 – 2010 was obtained to run

the model, however, the data is considered insufficient as the minimum

requirement is 30 observations. The reason researchers weren‟t able to get 30

years data was because China FDI inflows weren‟t available before the year 1982.

Secondly, hypothesis testing is inappropriate to be used in detecting

heteroscedasticity and autocorrelation problems. This is because the sample size

did not for the very least have 30 observations. For this reason, researchers use the

informal way which is graphical method to detect the heteroscedasticity and

autocorrelation problems. However, graphical method is a hard way to detect both

the problems because sometimes those problems are not obvious and subjective.

In addition, researchers will not be able to know the degree of heteroscedasticity

and autocorrelation problems from the graph. As a result, graphical method is

deemed an inappropriate method to detect the problems since it does not provide

any concrete result on the severity of the heteroscedasticity and autocorrelation

problems of the models.

Thirdly, the original model has multicollinearity problem so researchers weren‟t

able to run the model as a whole and interpret its result. Instead, researchers split

the model into 2 models and interpreted from there. Fourthly, even though

multicollinearity problem was solved, researchers still couldn‟t get the best result

due to heteroscedasticity and autocorrelation problems. Researchers were

incapable of solving both the heteroscedasticity and autocorrelation problem as

graphical method does not provide a concrete result of the heteroscedasticity and

autocorrelation problems in the model. In addition, the problems were way too

serious and not within the researchers‟ knowledge to solve it. The limitations are

acknowledged for it does not detract from the significance of findings but merely

provide platforms for future research.


Page 79 of 157

5.5 Recommendations for Future Research

Since sample size is the main root of the problems, it is highly recommended that

next researchers who are interested in further studying this paper should increase

the sample size to more than 30 observations. Researchers may use monthly,

quarterly or semiannual data instead of using annual data. This is because the

bigger the sample size, the lower the probability of having multicollinearity,

heteroscedasticity and autocorrelation problems. This will prevents the needs to

split the model but run it as a whole instead. Other than that, formal method like

hypothesis testing will be able to be applied on the detection of heteroscedasticity

and autocorrelation problems. Hypothesis testing will provides researchers with

better results in detecting the heteroscedasticity and autocorrelation problems.

And it will be made clear to researchers the severity of the problems, allowing

them to carry out the appropriate steps and solutions in solving the problems for

the best result.

The usage of annual data has always been bound with a higher chance of having

heteroscedasticity problem, so when using low frequency data like such, it is

recommended to increase the sample size. This will reduce the data frequency,

and hence, minimize the chances of getting heteroscedasticity problem. Besides,

future researchers are also advise to use other indicators for the independent

variable – quality of infrastructure - instead of the telephone line (per 100 people).

This is because telephone line (per 100 people) is not a good indicator for quality

of infrastructure. Last but not least, replacement of insignificant variable with

other relevant variables or adding in new variables like statutory corporate tax rate

and financial development will further improve the model. Or researchers may use

other methods such as generalized autoregressive conditional heteroscedasticity

(GRACH) method to run the data instead of multiple linear regression method.


Page 80 of 157

5.6 Conclusion

In conclusion, this paper proved that market size, economic growth, exchange rate,

trade openness, inflation rate and China FDI inflows are significant in determining

FDI inflows of Malaysia. Conversely, quality of infrastructure is determined as an

insignificant variable towards the Malaysia FDI inflows. All the results are

consistent and supported by the past research papers.

The results of this study can be a guideline and provide insight to policymakers

such as government and Bank Negara Malaysia in determining the ways to attract

more foreign direct investment inflow to Malaysia as foreign direct investment

inflows act as an important tool to enhance the economy of one‟s country. Besides,

this paper also assists practitioners such as investors and businessmen more

effectively in guiding and supporting the decision- making process either on

expanding markets or investment path.


Page 81 of 157

REFERENCES

Abdul Rahim, A.S. (2006). The changing role of FDI in the Malaysia economy-

an assessment. Retrieved May 12, 2012, from

http://www.facebook.com/l.php?u=http%3A%2F%2Feconomics.dstcentre.com

%2FThe%2520Changing%2520Role%2520Of%2520FDI%2520In%2520The

%2520Malaysian%2520Economy%2520By%2520Azmi%2520Shahrin.pdf&h

=HAQGYNkFt.

Addison, T. & Heshmati, A. (2003). The new global determinants of FDI flows to

developing countries: The importance of ICT and democratization. UNU World

Institute for Development Economics Research, Helsinki.

Ali, S. & Guo, W. (2005). Determinants of FDI in China. Journal of Global

Business and Technology, 1(2), 21-33.

Ang, J. B. (2008). Determinants of foreign direct investment in Malaysia. Journal

of Policy Modeling, 30, 185-189.

Aqeel, A. & Nishat, M. (2004). The determinants of foreign direct investment in

Pakistan. The Pakistan Development Review, 43(4), 651-664.

Artige, L. & Nicolini, R. (2005). Evidence on the determinants of foreign direct

investment: The case of three European regions. UFAE and IAE Working

papers, No. 655.05.

Asiedu, E. (2002). On the determinants of foreign direct investment to developing

countries: Is Africa different. World Development, 30(1), 107-119.

Athukorala, P. & Waglé, S. (2011). Foreign direct investment in Southeast Asia:

Is Malaysia falling behind. ASEAN Economic Bulletin, 28(2), 115-133.

Awan, M. Z., Khan, B., & Zaman, K. U. (2011). Economic determinants of

foreign direct investment (FDI) in commodity producing sector: A case study

of Pakistan. African Journal of Business Management, 5(2), 537-545.


Page 82 of 157

Azam, M. (2010). Economic determinants of foreign direct investment in Armenia,

Kyrgyz Republic, and Turkmenistan: Theory and evidence. Eurasian Journal

of Business and Economics, 3(6), 27-40.

Barrell, R. & Pain, N. (1996). An econometric analysis of U.S. foreign direct

investment. The Review of Economics and Statistics, 78(2), 200-207.

Benacek, V., Gronicki, M., Holland, D., & Sass, M. (2000). The determinants and

impact of foreign direct investment in Central and Eastern Europe. Journal of

United Nations, 9(3), 163-212.

Bevan, A. & Estrin, S. (2004). The determinants of foreign direct investment into

European transition economies. Journal of Comparative Economics, 32, 775-

787.

Busse, M. & Hefeker, C. (2007). Political risk, institutions and foreign direct

investment. European Journal of Political Economy, 23(2), 397-415.

Campa, J. M. (1993). Entry by foreign firms in the United States under exchange

rate uncertainty. The Review of Economics and Statistics, 75(4), 614-622.

Chakraborty, C. & Basu, P. (2002). Foreign direct investment and growth in India:

A cointegration approach. Applied Economics, 34, 1061-1073.

Chantasasawat, B., Fung, K. C., Iizaka, H., & Siu, A. (2010). FDI flows to Latin

America, East and Southeast Asia, and China: Substitutes or complements.

Review of Development Economics, 14(3), 533-546.

Chantasasawat, B., Fung, K.C., Iizaka, H., & Siu, A. (2004). The giant sucking

sound: Is China diverting foreign direct investment from other Asian

economies. Asian Economic Papers, 3(3), 122-144.

Choe, J. I. (2003). Do foreign direct investment and gross domestic investment

promote economic growth. Review of Development Economics, 7(1), 44-57.

Demirhan, E. & Masca, M. (2008). Determinants of foreign direct investment

flows to developing countries: A cross-sectional analysis. Prague Economic

Papers, 4, 356-369.


Page 83 of 157

Dunning, J.H. (1995). Reappraising the eclectic paradigm in an age of alliance

capitalism. Journal of International Business Studies, 26(3), 461-491.

Eichengreen, B. & Tong, H. (2006). Fear of China. Journal of Asian Economics,

17, 226-240.

Eichengreen, B. & Tong, H. (2007). Is China‟s FDI coming at the expense of

other countries? Journal of the Japanese and International Economies, 21(2),

153-172.

Fan, P.H., Morck, R., Xu, L.C., & Yeung, B. (2007). Does „good government‟

draw foreign capote: Explaining China‟s exceptional foreign direct investment

inflow. World Bank Policy Research Working Paper. Retrieved 12 May, 2012

from http://elibrary.worldbank.org/content/workingpaper/10.1596/1813-9450-4206

Froot, K. A. & Stein, J.C. (1991). Exchange rates and foreign direct investment:

An imperfect capital markets approach. Quarterly Journal of Economics,

106(4), 1191-1217.

Gao, T. (2005). Foreign direct investment and growth under economic integration.

Journal of international economics, 67, 157-174.

Globerman, S. & Shapiro, D. (2003). Governance infrastructure and US foreign

direct investment. Journal of International Business Studies, 34(1), 19-39.

Goodspeed, T., Martinez-Vazquez, J., & Zhang, L. (2006). Are other government

policies more important than taxation in attracting FDI. Andrew Young School

of Policy Studies, Working Paper 06-28.

Gujarati, D.N. & Porter, D.C. (2009). Basic econometrics. (5th ed.). Boston, MA:

McGraw-Hill/Irwin.

Guo, J. Q. & Trivedi, P. K. (2002). Firm-specific assets and the link between

exchange rates and Japanese foreign direct investment in the United States: A

re-examination. The Japanese Economic Review, 53(3), 337-349.


Page 84 of 157

Hansen, H. & Rand, J. (2006). On the causal links between FDI and growth in

developing countries. The World Economy, 29(1), 21-41.

Hara, M. & Razafimahefa, I. F. (2005). The determinants of foreign direct

investments into Japan. Kobe University Economic Review, 51, 21-34.

Kahai, S. K. (2011). Traditional and non-traditional determinants of foreign direct

investment in developing countries. Journal of Applied Business Research,

20(1), 43-50.

Karimi, M.S., Yusop, Z., & Law, S.H. (2010). Location decision for foreign direct

investment in ASEAN countries: A TOPSIS approach. International Research

Journal of Finance and Economics, (36), 2010.

Kiyota, K. & Urata, S. (2004). Exchange rate, exchange rate volatility and foreign

direct investment. The World Economy, 27(10), 1501-1536.

Klein, M.W. & Rosengren, E. (1994). The real exchange rate and foreign direct

investment in the United States: Relative wealth vs. relative wage effects.

Journal of International Economics, 36(3), 373-379.

Kolstad, I. & Villanger, E. (2008). Determinants of foreign direct investment in

services. European Journal of Political Economy, 24(2), 518-533.

Larget, B. (2007, February 20). Model Selection and Multicollinearity. pp. 1-4.

Lim, S.H. (2008). How investment promotion affects attracting foreign direct

investment: Analytical argument and empirical analyses. International

Business Review, 17, 39-53.

Loree, D.W. & Guisinger, S. (1995). Policy and non-policy determinant of US

equity foreign direct investment. Journal of Business Studies, 26(2), 281-299.

Lucey, T. (2002). Quantitative techniques. Cheriton, NW: Cengage Learning

EMEA.


Page 85 of 157

Luiz, J. M. & Charalambous, H. (2009). Factors influencing foreign direct

investment of South African financial services firms in Sub-Saharan Africa.

International Business Review, 18, 305-317.

McDaniel, C. & Gates, R. (2010). Marketing Research. (8th ed.). Hoboken, NJ:

John Wiley & Sons.

Medvedev, D. (2012). Beyond trade: The impact of preferential trade agreements

on FDI inflows. World Development, 40(1), 49-61.

Mengistu, A.A. & Adhikary, B. K. (2011). Does good governance matter for FDI

inflows? Evidence from Asian economies. Asia Pacific Business Review, 17(3),

281-299.

Moffett, M.H., Stonehill, A.I. & Eitheman, D.K. (2009). Fundamentals of

Multinational Finance. (3th ed.). Boston: Addison Wesley.

Moosa, I. A. & Cardak, B. A. (2006). The determinants of foreign direct

investment: An extreme bounds analysis. Journal of Multinational Financial

Management, 16(2), 199-211.

Moosa, I. A. (2002). Foreign direct investment: theory, evidence, and practice.

New York, NY: Palgave Macmillan.

Morisset, J. (2000). Foreign direct investment in Africa policies also matter.

Transnational Corporations, 9(2), 281-289.

Motulsky, H. (2002). Multicollinearity in multiple regression. Retrieved March 1,

2012, from GraphPad.com:

http://www.graphpad.com/articles/Multicollinearity.htm

Nurudeen, A. & Wafure, O.G. (2010). Determinants of foreign direct investment

in Nigeria: An empirical analysis. Global Journal of Human Social Science,

10(1), 26-34.

Nurudeen, A. , Wafure, O. G., & Auta, E. M. (2011). Determinants of foreign

direct investment: The case of Nigeria. The IUP Journal of Monetary

Economics, 9(3), 50-68.


Page 86 of 157

Quazi, R. (2007). Economic freedom and foreign direct investment in East Asia.

Journal of the Asia Pacific Economy, 12(3), 329-344.

Quer, D. & Claver, E. (2007). Determinants of Spanish foreign direct investment

in Morocco. Emerging Markets Finance and Trade, 43(2), 19-32.

Rehman, C. A., Ilyas, M., Alam, H. M., & Akram, M. (2011). The impact of

infrastructure on foreign direct investment: The case of Pakistan. International

Journal of Business and Management, 6(5), 268-276.

Renfro, C. G. (2004). A compendium of existing econometric software packages.

Journal of Economic and Social Measurement , 29, 359-409.

Rodriguez, X.A. & Pallas, J. (2008). Determinants of foreign direct investment in

Spain. Applied Economics, 40, 2443-2450.

Salike, N. (2010). Investigating of the “China effect” on crowding out Japanese

FDI: An industry-level analysis. China Economic Review, 21(4), 582-597.

Sawkut, R., Boopen, S., Taruna, R., & Vinesh, S. (2009). Determinants of FDI:

Lessons from African economies. Journal of Applied Business and Economics,

9(1), 293-308.

Seim, L. T. (2009). FDI and openness: Differences in response across countries.

Chr. Michelsen Institute, Working Paper.

Shamsuddin, A.F.M. (1994). Economic determinants of foreign direct investment

in less developed countries. The Pakistan Development Review, 33(1), 41-51.

Sharma, K. & Bandara, Y. (2010). Trends, patterns and determinants of Australian

foreign direct investment. Journal of Economic Issues, 64(3), 661-676.

Srinivasan, P. (2011). Determinants of foreign direct investment in SAARC

nations: An econometric investigation. IUP Journal of Managerial Economics,

9(3), 26-47.


Page 87 of 157

Startz, R. (2007). Eviews illustrated for version 7. Quantitative Micro Software .

Sukamolson, S. (2005). Fundamentals of quantitative research. Retrieved March

3, 2012, from http://www.culi.chula.ac.th/e-

Journal/bod/Suphat%20Sukamolson.pdf

The economic watch. (2010). Definition of foreign direct investment (fdi).

Retrieved March 3, 2012, from http://www.economywatch.com/foreign-direct-

investment/definition.html

Trevino, L. J. & Mixon Jr, F.G. (2004). Strategic factors affecting foreign direct

investment decisions by multi-national enterprises in Latin America. Journal of

World Business, 39, 233-243.

Vadlamannati, K. C., Tamazian, A., & Irala, L. R. (2009). Determinants of foreign

direct investment and volatility in South East Asian economies. Journal of the

Asia Pacific Economy, 14(3), 246-261.

Vijayakumar, N., Sridharan, P., & Rao, K.C.S. (2010). Determinants of FDI in

BRICS countries: A panel analysis. Int. Journal of Business Science and

Applied Management, 5(3), 1-13.

Wang, C.G., Wei, Y.Q., & Liu, X.M. (2007). Does China rival its neighbouring

economies for inward FDI? Transnational Corporations, 16(3), 35-60.

Wheeler, D. & Mody, A. (1992). International investment location decisions: The

case of U.S. firms. Journal of International Economics, 33, 57-76.

Wong, H.K. & Jomo, K.S. (2005). Before the storm: The impact of foreign capital

inflows on the Malaysian economy, 1966-1996. Journal of the Asia Pacific

Economy, 10(1), 56-69.

Zhang, H.L. (2001). Does foreign direct investment promote economic growth?

Evidence from East Asian and Latin America. Comtemporary Economic Policy,

19(2), 175-185.


Page 88 of 157

Appendices

Appendix 1:

Empirical Result of Multiple Linear Regression Model

Dependent Variable: MFDI

Method: Least Squares

Date: 06/03/12 Time: 18:03

Sample: 1982 2010

Included observations: 29

Coefficient Std. Error t-Statistic Prob.

C 2.00E+09 2.09E+09 0.956468 0.3497

GDP -0.010788 0.021409 -0.503912 0.6196

GDPG 1.53E+08 86462339 1.767404 0.0917

OREER -2.32E+09 9.89E+08 -2.344776 0.0289

TL 8122953. 1.93E+08 0.042122 0.9668

TO 3.78E+09 2.94E+09 1.286072 0.2124

INF 14902615 1.26E+08 0.118276 0.9070

CFDI 0.044008 0.021564 2.040837 0.0540

R-squared 0.832961 Mean dependent var 3.46E+09

Adjusted R-squared 0.777282 S.D. dependent var 2.45E+09

S.E. of regression 1.16E+09 Akaike info criterion 44.80748

Sum squared resid 2.82E+19 Schwarz criterion 45.18466

Log likelihood -641.7084 Hannan-Quinn criter. 44.92561

F-statistic 14.95993 Durbin-Watson stat 1.836526

Prob(F-statistic) 0.000001


Page 89 of 157

Appendix 2:

Normality Test of Multiple Linear Regression Model (Jarque-

Bera Normality Test)

H0 : Error terms are normally distributed

H1: Error terms are not normally distributed

Critical value: α= 0.10

Test statistic: p-value = 0.790615

Decision rules: Reject H0 if p-value less than α= 0.10, otherwise do not

reject H0.

Decision: Do not reject H0, since p-value (0.790615) is more than α= 0.10

Conclusion: We have enough evidence to conclude that the error terms are

normally distributed.

0

1

2

3

4

5

6

7

8

9

-2.0e+09 -1.0e+09 0.00000 1.0e+09 2.0e+09

Series: Residuals

Sample 1982 2010

Observations 29

Mean -1.79e-06

Median -25502905

Maximum 2.04e+09

Minimum -1.90e+09

Std. Dev. 1.00e+09

Skewness -0.162377

Kurtosis 2.467640

Jarque-Bera 0.469888

Probability 0.790615


Page 90 of 157

Appendix 3:

Model Specification Test of Multiple Linear Regression Model

(Ramsey RESET Test)

Ramsey RESET Test:

F-statistic 5.651726 Prob. F(1,20) 0.0275

Log likelihood ratio 7.217479 Prob. Chi-Square(1) 0.0072

Test Equation:

Dependent Variable: MFDI


Date: 06/03/12 Time: 18:08

Sample: 1982 2010



C -3.82E+09 3.09E+09 -1.235359 0.2310

GDP 0.028093 0.025352 1.108120 0.2810

GDPG 25792707 94735647 0.272260 0.7882

OREER 1.57E+09 1.86E+09 0.842028 0.4097

TL -38062711 1.76E+08 -0.216807 0.8306

TO -3.03E+08 3.17E+09 -0.095667 0.9247

INF 1.17E+08 1.22E+08 0.959021 0.3490

CFDI -0.054719 0.045883 -1.192562 0.2470

FITTED^2 1.83E-10 7.69E-11 2.377336 0.0275








H0: Model specification is correct.

H1: Model specification is incorrect.

Critical Value: α = 0.10

p-value = 0.0275

Decision rules: Reject H0, if p-value less than α = 0.10, otherwise do not

reject H0.

Decision: Reject H0, since the p-value (0.0275) less than the significance

level, 0.10.

Conclusion: We have enough evidence to conclude that the model

specification is incorrect.


Page 91 of 157

Appendix 4:

Multicollinearity Testing of Multiple Linear Regression Model

4.1 Correlation table

CFDI GDP GDPG INF OREER TL TO

CFDI 1.000000 0.971965 -0.136784 -0.194607 0.547252 0.553830 0.557112

GDP 0.971965 1.000000 -0.102220 -0.196632 0.576916 0.640295 0.615826

GDPG -0.136784 -0.102220 1.000000 0.026234 -0.392392 -0.145708 -0.046531

INF -0.194607 -0.196632 0.026234 1.000000 -0.313685 -0.234635 -0.198727

OREER 0.547252 0.576916 -0.392392 -0.313685 1.000000 0.797065 0.828626

TL 0.553830 0.640295 -0.145708 -0.234635 0.797065 1.000000 0.953759

TO 0.557112 0.615826 -0.046531 -0.198727 0.828626 0.953759 1.000000

4.2 Variance inflation factor (VIF) table

Independent variables Correlation, R R2 VIF

CFDI GDP 0.971965 0.944715961 18.08840342

CFDI GDPG -0.136784 0.018709862 1.019066595

CFDI INF -0.194607 0.037871884 1.039362621

CFDI OREER 0.547252 0.299484751 1.427520673

CFDI TL 0.553830 0.306727668 1.442434602

CFDI TO 0.557112 0.310373780 1.450060875

GDP GDPG -0.102220 0.010448928 1.010559261

GDP INF -0.196632 0.038664143 1.040219183

GDP OREER 0.576916 0.332832071 1.498873007

GDP TL 0.640295 0.409977687 1.694851157

GDP TO 0.615826 0.379241662 1.610932852

GDPG INF 0.026234 0.000688223 1.000688697

GDPG OREER -0.392392 0.153971481 1.181993251

GDPG TL -0.145708 0.021230821 1.021691346

GDPG TO -0.046531 0.002165134 1.002169832

INF OREER -0.313685 0.098398279 1.109137191

INF TL -0.234635 0.055053583 1.058261063

INF TO -0.198727 0.039492420 1.041116198

OREER TL 0.797065 0.635312614 2.742074550

OREER TO 0.828626 0.686621047 3.191024765

TL TO 0.953759 0.909656230 11.06883186


Page 92 of 157

Appendixes 5:

Heteroscedasticity Testing (Autoregressive Conditional

Heteroscedasticity (ARCH) Test)

5.1 Lag length = 1

Heteroskedasticity Test: ARCH


Obs*R-squared 0.025103 Prob. Chi-Square(1) 0.8741

Test Equation:

Dependent Variable: RESID^2


Date: 06/04/12 Time: 18:06

Sample (adjusted): 1983 2010

Included observations: 28 after adjustments


C 1.03E+18 2.96E+17 3.496311 0.0017

RESID^2(-1) -0.030139 0.197319 -0.152744 0.8798


Adjusted R-squared -0.037531 S.D. dependent var 1.21E+18







Page 93 of 157

5.2 Lag length = 2




Test Equation:



Date: 06/04/12 Time: 18:07




C 1.45E+18 3.53E+17 4.121318 0.0004

RESID^2(-1) -0.096185 0.193300 -0.497595 0.6233

RESID^2(-2) -0.380137 0.208202 -1.825812 0.0803









Page 94 of 157

5.3 Lag length = 3




Test Equation:



Date: 06/04/12 Time: 18:07




C 1.48E+18 4.79E+17 3.081805 0.0055

RESID^2(-1) -0.108133 0.214061 -0.505154 0.6185

RESID^2(-2) -0.404152 0.218537 -1.849356 0.0779

RESID^2(-3) 0.053557 0.228521 0.234364 0.8169









Page 95 of 157

5.4 Lag length = 4




Test Equation:



Date: 06/04/12 Time: 18:08




C 1.31E+18 5.95E+17 2.203757 0.0394

RESID^2(-1) -0.135413 0.222285 -0.609188 0.5493

RESID^2(-2) -0.356580 0.241090 -1.479034 0.1547

RESID^2(-3) 0.044416 0.240053 0.185026 0.8551

RESID^2(-4) 0.208781 0.234555 0.890112 0.3840









Page 96 of 157

5.5 Lag length = 5




Test Equation:



Date: 06/04/12 Time: 18:08




C 1.50E+18 6.91E+17 2.163414 0.0442

RESID^2(-1) -0.140870 0.241419 -0.583511 0.5668

RESID^2(-2) -0.376745 0.259611 -1.451189 0.1639

RESID^2(-3) -0.005722 0.265627 -0.021543 0.9830

RESID^2(-4) 0.170954 0.250522 0.682391 0.5037

RESID^2(-5) -0.050662 0.262107 -0.193287 0.8489









Page 97 of 157

5.6 Lag length = 6




Test Equation:



Date: 06/04/12 Time: 18:08




C 1.50E+18 8.25E+17 1.814958 0.0883

RESID^2(-1) -0.129595 0.258090 -0.502131 0.6224

RESID^2(-2) -0.396028 0.270229 -1.465530 0.1622

RESID^2(-3) -0.091170 0.291319 -0.312957 0.7584

RESID^2(-4) 0.190979 0.281756 0.677819 0.5076

RESID^2(-5) -0.107001 0.278654 -0.383993 0.7060

RESID^2(-6) 0.196640 0.323161 0.608489 0.5514









Page 98 of 157

5.7 Summary of Heteroscedasticity Testing, Autoregressive Conditional

Heteroscedasticity (ARCH) test result

Lag length 2 has the lowest AIC

H0: There is no heteroscedasticity problem.

H1: There is heteroscedasticity problem.

Critical Value: 0.10

Decision rule: Reject H0 if the p-value is smaller than the critical value.

Otherwise, do not reject H0.

P-value: 0.2025

Decision: Do not reject H0 since the p-value, 0.2025 is greater than the

critical value, 0.10.

Conclusion: There is not enough evidence to conclude that the model

has heteroscedasticity problem in the estimated model at 10%

significant level.

Lag Length AIC p-value

1 86.21031 0.8798

2 86.16852 0.2025

3 86.25941 0.3159

4 86.33848 0.3603

5 86.45429 0.5273

6 86.55678 0.6211


Page 99 of 157

Appendix 6:

Autocorrelation Testing (Breusch-Godfrey LM Test)

6.1 Lag length = 1

Breusch-Godfrey Serial Correlation LM Test:



Test Equation:

Dependent Variable: RESID


Date: 06/04/12 Time: 17:49

Sample: 1982 2010


Presample missing value lagged residuals set to zero.


C 29637137 2.14E+09 0.013847 0.9891

GDP 0.000767 0.022207 0.034556 0.9728

GDPG 395466.3 88516595 0.004468 0.9965

OREER -13156507 1.01E+09 -0.012976 0.9898

TL -1218482. 1.97E+08 -0.006171 0.9951

TO -2485350. 3.01E+09 -0.000826 0.9993

INF -4224690. 1.30E+08 -0.032378 0.9745

CFDI -0.000533 0.022213 -0.024016 0.9811

RESID(-1) 0.052462 0.246204 0.213083 0.8334

R-squared 0.002265 Mean dependent var -1.79E-06








Page 100 of 157

6.2 Lag length = 2




Test Equation:



Date: 06/04/12 Time: 17:50

Sample: 1982 2010




C -4.44E+08 2.10E+09 -0.211480 0.8348

GDP 0.006162 0.021851 0.282011 0.7810

GDPG -13330719 86392656 -0.154304 0.8790

OREER -2.06E+08 9.92E+08 -0.207341 0.8379

TL -67079349 1.97E+08 -0.341143 0.7367

TO 1.10E+09 3.01E+09 0.363694 0.7201

INF -10720089 1.27E+08 -0.084609 0.9335

CFDI -0.006284 0.021898 -0.286982 0.7772

RESID(-1) 0.109489 0.241961 0.452507 0.6560

RESID(-2) -0.381544 0.255198 -1.495088 0.1513









Page 101 of 157

6.3 Lag length = 3




Test Equation:



Date: 06/04/12 Time: 17:51

Sample: 1982 2010




C 3.70E+08 1.95E+09 0.190166 0.8513

GDP -0.010203 0.021206 -0.481116 0.6362

GDPG 40176333 82229255 0.488589 0.6310

OREER -3.17E+08 9.05E+08 -0.350287 0.7302

TL 55884566 1.87E+08 0.298351 0.7689

TO 9509198. 2.78E+09 0.003415 0.9973

INF 17600759 1.16E+08 0.151713 0.8811

CFDI 0.009888 0.021216 0.466047 0.6468

RESID(-1) -0.098269 0.239229 -0.410776 0.6861

RESID(-2) -0.229834 0.242079 -0.949418 0.3550

RESID(-3) -0.606058 0.272701 -2.222430 0.0393









Page 102 of 157

6.4 Lag length = 4




Test Equation:



Date: 06/04/12 Time: 17:51

Sample: 1982 2010




C 9.51E+08 2.08E+09 0.456757 0.6536

GDP -0.012994 0.021638 -0.600511 0.5561

GDPG 49720925 83682212 0.594164 0.5602

OREER -5.62E+08 9.58E+08 -0.587197 0.5648

TL 1.02E+08 1.97E+08 0.517858 0.6112

TO -1.83E+08 2.82E+09 -0.064887 0.9490

INF 4410228. 1.18E+08 0.037369 0.9706

CFDI 0.012072 0.021548 0.560247 0.5826

RESID(-1) -0.227795 0.286230 -0.795844 0.4371

RESID(-2) -0.273597 0.249570 -1.096277 0.2882

RESID(-3) -0.642857 0.278416 -2.308981 0.0338

RESID(-4) -0.273448 0.325341 -0.840496 0.4123









Page 103 of 157

6.5 Lag length = 5




Test Equation:



Date: 06/04/12 Time: 17:52

Sample: 1982 2010




C 1.39E+09 1.61E+09 0.862072 0.4014

GDP 0.009345 0.017890 0.522350 0.6086

GDPG -42227512 69762269 -0.605306 0.5535

OREER -1.36E+09 7.75E+08 -1.761726 0.0972

TL 95263671 1.52E+08 0.626604 0.5398

TO 9.44E+08 2.20E+09 0.428687 0.6739

INF 8913623. 91264692 0.097668 0.9234

CFDI -0.014602 0.018299 -0.797970 0.4366

RESID(-1) -0.195390 0.221512 -0.882072 0.3908

RESID(-2) -0.865631 0.255791 -3.384134 0.0038

RESID(-3) -0.737953 0.216962 -3.401302 0.0037

RESID(-4) -0.273482 0.251563 -1.087132 0.2931

RESID(-5) -0.991152 0.281087 -3.526138 0.0028









Page 104 of 157

6.6 Lag length = 6




Test Equation:



Date: 06/04/12 Time: 17:52

Sample: 1982 2010




C 1.57E+09 1.77E+09 0.885265 0.3900

GDP 0.008702 0.018560 0.468885 0.6459

GDPG -43213646 71931980 -0.600757 0.5570

OREER -1.45E+09 8.54E+08 -1.700501 0.1097

TL 1.10E+08 1.65E+08 0.667966 0.5143

TO 9.18E+08 2.27E+09 0.404313 0.6917

INF 5331471. 94812440 0.056232 0.9559

CFDI -0.014179 0.018903 -0.750087 0.4648

RESID(-1) -0.236763 0.269422 -0.878782 0.3934

RESID(-2) -0.886242 0.272951 -3.246888 0.0054

RESID(-3) -0.810750 0.336925 -2.406323 0.0295

RESID(-4) -0.323944 0.312543 -1.036478 0.3164

RESID(-5) -1.008321 0.295548 -3.411703 0.0039

RESID(-6) -0.097382 0.337323 -0.288692 0.7768









Page 105 of 157

6.7 Summary of Autocorrelation Testing, Breusch-Godfrey LM test result

Lag length 5 has the minimum AIC

Ho: There is no autocorrelation.

H1: There is autocorrelation.

Critical value: = 0.10

Decision rule: Reject Ho. If the p-value is less than the significant level,

= 0.10. Otherwise, do not reject Ho.

P-Value = 0.0048

Decision making: Reject Ho. Since the p-value 0.0048 is smaller than the

significant level at 0.10.

Conclusion: There is sufficient evidence to conclude that there is

autocorrelation problem in the estimated model at 10% significant level.


1 44.87418 0.8334

2 44.83192 0.3402

3 44.65841 0.0867

4 44.68666 0.1299

5 44.18064 0.0048

6 44.24406 0.0119


Page 106 of 157

Appendix 7:

Residual Graph of the Multiple Linear Regression Model

-2,000,000,000

-1,000,000,000

0

1,000,000,000

2,000,000,000

3,000,000,000

1985 1990 1995 2000 2005 2010

MFDI Residuals


Page 107 of 157

Appendix 8:

Empirical Result of Semi-logarithmic: Log-Lin Model

Dependent Variable: LOG(MFDI)


Date: 06/15/12 Time: 17:27

Sample: 1982 2010



C 19.37893 0.814918 23.78021 0.0000

GDP 2.91E-12 8.36E-12 0.347688 0.7315

GDPG 0.054179 0.033766 1.604566 0.1235

OREER -0.803537 0.386136 -2.080971 0.0499

TL -0.056388 0.075309 -0.748748 0.4623

TO 2.654155 1.148582 2.310810 0.0311

INF 0.066051 0.049206 1.342345 0.1938

CFDI 4.89E-12 8.42E-12 0.580837 0.5675

R-squared 0.813519 Mean dependent var 21.64097

Adjusted R-squared 0.751359 S.D. dependent var 0.907256

S.E. of regression 0.452394 Akaike info criterion 1.480423

Sum squared resid 4.297861 Schwarz criterion 1.857608





Page 108 of 157

Appendix 9:

Normality Test of Semi-logarithmic: Log-Lin Model (Jarque-

Bera Normality Test)






reject H0.




0

1

2

3

4

5

6

7

8

9

-1.0 -0.5 0.0 0.5

Series: Residuals

Sample 1982 2010

Observations 29

Mean -1.12e-15

Median 0.047150

Maximum 0.663597

Minimum -1.097911

Std. Dev. 0.391784

Skewness -0.654396

Kurtosis 3.521145




Page 109 of 157

Appendix 10:

Model Specification test of Semi-Logarithmic: Log-Lin Model

(Ramsey RESET Test)

Ramsey RESET Test:



Test Equation:



Date: 06/15/12 Time: 17:42

Sample: 1982 2010

Included observations: 29 Coefficient Std. Error t-Statistic Prob.

C 64.07198 67.27116 0.952444 0.3522

GDP 1.58E-11 2.12E-11 0.746192 0.4642

GDPG 0.322988 0.406020 0.795497 0.4357

OREER -5.059446 6.417378 -0.788398 0.4397

TL -0.346041 0.442579 -0.781872 0.4434

TO 16.34850 20.64378 0.791934 0.4377

INF 0.388013 0.487134 0.796522 0.4351

CFDI 3.27E-11 4.27E-11 0.765540 0.4529

FITTED^2 -0.119100 0.179254 -0.664422 0.5140











p-value = 0.5140


reject H0.

Decision: Reject H0, since the p-value (0.5140) more than the significance

level, 0.10


specification is correct.


Page 110 of 157

Appendix 11:

Empirical Result of Model 1



Date: 06/15/12 Time: 18:28

Sample: 1982 2010


C 19.62773 0.705766 27.81055 0.0000

GDPG 0.068305 0.026728 2.555526 0.0177

OREER -0.693089 0.346083 -2.002668 0.0571

TO 1.860519 0.461854 4.028367 0.0005

INF 0.077766 0.045178 1.721305 0.0986

CFDI 7.66E-12 1.84E-12 4.169426 0.0004









Page 111 of 157

Appendix 12:




Date: 06/15/12 Time: 18:30

Sample: 1982 2010


C 19.65046 0.821058 23.93310 0.0000

GDPG 0.106277 0.026020 4.084340 0.0005

OREER -0.172537 0.308930 -0.558499 0.5819

TL 0.087694 0.030484 2.876740 0.0085

INF 0.107534 0.049503 2.172258 0.0404

CFDI 7.92E-12 2.06E-12 3.843187 0.0008









Page 112 of 157

Appendix 13:




Date: 06/15/12 Time: 18:36

Sample: 1982 2010


C 19.46177 0.723421 26.90242 0.0000

GDP 7.04E-12 1.84E-12 3.830467 0.0009

GDPG 0.065417 0.027658 2.365210 0.0268

OREER -0.673539 0.358007 -1.881360 0.0726

TO 1.773820 0.484693 3.659679 0.0013

INF 0.076571 0.046766 1.637319 0.1152









Page 113 of 157

Appendix 14:




Date: 06/15/12 Time: 18:37

Sample: 1982 2010


C 19.35685 0.851557 22.73113 0.0000

GDP 7.17E-12 2.14E-12 3.343553 0.0028

GDPG 0.103847 0.027389 3.791622 0.0009

OREER -0.104738 0.322322 -0.324948 0.7482

TL 0.075415 0.033241 2.268752 0.0330

INF 0.105715 0.052028 2.031886 0.0539









Page 114 of 157

Appendix 15:

Normality Test of Model 1 (Jarque-Bera Normality Test)






reject H0.




0

1

2

3

4

5

6

7

8

9

-1.0 -0.5 0.0 0.5

Series: Residuals

Sample 1982 2010

Observations 29

Mean 5.63e-16

Median 0.058606

Maximum 0.677370

Minimum -1.138448

Std. Dev. 0.397159

Skewness -0.633643

Kurtosis 3.731610




Page 115 of 157

Appendix 16:

Model Specification Test of Model 1 (Ramsey RESET Test)

Ramsey RESET Test:



Test Equation:



Date: 06/15/12 Time: 19:13

Sample: 1982 2010


C 44.21453 63.59844 0.695214 0.4942

GDPG 0.250366 0.471692 0.530782 0.6009

OREER -2.668913 5.122673 -0.521000 0.6076

TO 6.991600 13.28001 0.526475 0.6038

INF 0.282478 0.531491 0.531482 0.6004

CFDI 2.90E-11 5.53E-11 0.524720 0.6050

FITTED^2 -0.063731 0.164842 -0.386619 0.7028











p-value = 0.7028


reject H0.

Decision: Do not reject H0, since the p-value (0.7028) more than the

significance level, 0.10


specification is correct.


Page 116 of 157

Appendix 17:

Multicollinearity Testing of Model

17.1 Correlation table

GDPG INF OREER TL CFDI

GDPG 1.000000 0.026234 -0.392392 -0.145708 -0.136784

INF 0.026234 1.000000 -0.313685 -0.234635 -0.194607

OREER -0.392392 -0.313685 1.000000 0.797065 0.547252

TL -0.145708 -0.234635 0.797065 1.000000 0.553830

CFDI -0.136784 -0.194607 0.547252 0.553830 1.000000


Page 117 of 157

Appendix 18:


Heteroscedasticity (ARCH) Test) of Model 1

18.1 Lag length = 1




Test Equation:



Date: 06/15/12 Time: 19:25


Included observations: 28 after adjustments Coefficient Std. Error t-Statistic Prob.

C 0.170728 0.058071 2.939985 0.0068

RESID^2(-1) -0.088992 0.194443 -0.457676 0.6510


Adjusted R-squared -0.030162 S.D. dependent var 0.259561







Page 118 of 157

18.2 Lag length = 2




Test Equation:



Date: 06/15/12 Time: 19:25



C 0.221140 0.067304 3.285675 0.0031

RESID^2(-1) -0.124215 0.195557 -0.635187 0.5313

RESID^2(-2) -0.258310 0.195187 -1.323397 0.1982









Page 119 of 157

18.3 Lag length = 3




Test Equation:



Date: 06/15/12 Time: 19:26



C 0.181778 0.082977 2.190714 0.0394

RESID^2(-1) -0.077363 0.206746 -0.374195 0.7118

RESID^2(-2) -0.245480 0.199806 -1.228594 0.2322

RESID^2(-3) 0.225106 0.205434 1.095758 0.2850









Page 120 of 157

18.4 Lag length = 4




Test Equation:



Date: 06/15/12 Time: 19:26



C 0.218762 0.096073 2.277040 0.0339

RESID^2(-1) -0.047571 0.219787 -0.216441 0.8308

RESID^2(-2) -0.294433 0.213769 -1.377344 0.1836

RESID^2(-3) 0.205367 0.214028 0.959535 0.3487

RESID^2(-4) -0.169736 0.218473 -0.776917 0.4463









Page 121 of 157

18.5 Lag length = 5




Test Equation:



Date: 06/15/12 Time: 19:27



C 0.265263 0.115383 2.298984 0.0337

RESID^2(-1) -0.082071 0.231756 -0.354125 0.7274

RESID^2(-2) -0.284342 0.227887 -1.247732 0.2281

RESID^2(-3) 0.145089 0.234405 0.618967 0.5437

RESID^2(-4) -0.193102 0.227749 -0.847871 0.4076

RESID^2(-5) -0.125086 0.233445 -0.535827 0.5986









Page 122 of 157

18.6 Lag length = 6




Test Equation:



Date: 06/15/12 Time: 19:27



C 0.297861 0.144331 2.063741 0.0557

RESID^2(-1) -0.096138 0.248516 -0.386850 0.7040

RESID^2(-2) -0.307895 0.243978 -1.261978 0.2250

RESID^2(-3) 0.173575 0.252197 0.688251 0.5012

RESID^2(-4) -0.230216 0.251478 -0.915451 0.3735

RESID^2(-5) -0.131756 0.249184 -0.528749 0.6042

RESID^2(-6) -0.146310 0.248708 -0.588280 0.5646









Page 123 of 157


Heteroscedasticity (ARCH) test result of Model 1






Otherwise,

do not reject H0.

P-value: 0.6510



Conclusion: There is not enough evidence to conclude that the model has

heteroscedasticity problem in the estimated model at 10% significant level.


1 0.238813 0.6510

2 0.268065 0.3785

3 0.326720 0.3634

4 0.422529 0.4563

5 0.524141 0.5713

6 0.652611 0.7024


Page 124 of 157

Appendix 19:

Autocorrelation Testing (Breusch-Godfrey LM Test) of Model

1

19.1 Lag length = 1




Test Equation:



Date: 06/15/12 Time: 19:41

Sample: 1982 2010


Presample missing value lagged residuals set to zero. Coefficient Std. Error t-Statistic Prob.

C 0.043871 0.723153 0.060667 0.9522

GDPG 0.000547 0.027202 0.020100 0.9841

OREER -0.001928 0.351940 -0.005478 0.9957

TO -0.017321 0.470953 -0.036778 0.9710

INF -0.005739 0.047390 -0.121099 0.9047

CFDI 1.35E-13 1.89E-12 0.071767 0.9434

RESID(-1) 0.111358 0.225674 0.493446 0.6266

R-squared 0.010947 Mean dependent var 5.63E-16








Page 125 of 157

19.2 Lag length = 2




Test Equation:



Date: 06/15/12 Time: 19:42

Sample: 1982 2010



C 0.056852 0.735815 0.077264 0.9391

GDPG 0.001334 0.027703 0.048154 0.9620

OREER -0.036946 0.364106 -0.101471 0.9201

TO 0.033483 0.488675 0.068518 0.9460

INF -0.005173 0.048204 -0.107309 0.9156

CFDI 2.07E-13 1.92E-12 0.107688 0.9153

RESID(-1) 0.128084 0.231711 0.552774 0.5863

RESID(-2) -0.125429 0.239796 -0.523066 0.6064









Page 126 of 157

19.3 Lag length = 3




Test Equation:



Date: 06/15/12 Time: 19:42

Sample: 1982 2010



C 0.412683 0.597754 0.690389 0.4979

GDPG -0.001474 0.022204 -0.066366 0.9477

OREER -0.376291 0.306759 -1.226667 0.2342

TO 0.412342 0.405567 1.016705 0.3214

INF 0.002579 0.038672 0.066694 0.9475

CFDI 9.99E-13 1.56E-12 0.641644 0.5284

RESID(-1) 0.034641 0.187436 0.184817 0.8552

RESID(-2) -0.072011 0.192656 -0.373781 0.7125

RESID(-3) -0.720574 0.201946 -3.568162 0.0019









Page 127 of 157

19.4 Lag length = 4




Test Equation:



Date: 06/15/12 Time: 19:42

Sample: 1982 2010



C 0.542814 0.632951 0.857592 0.4018

GDPG -0.005118 0.023075 -0.221806 0.8268

OREER -0.500468 0.357176 -1.401181 0.1773

TO 0.567850 0.466280 1.217831 0.2382

INF 0.005147 0.039337 0.130847 0.8973

CFDI 1.04E-12 1.58E-12 0.656828 0.5192

RESID(-1) -0.078440 0.248563 -0.315575 0.7558

RESID(-2) -0.099942 0.199111 -0.501942 0.6215

RESID(-3) -0.730766 0.205046 -3.563913 0.0021

RESID(-4) -0.190818 0.270750 -0.704773 0.4895









Page 128 of 157

19.5 Lag length = 5




Test Equation:



Date: 06/15/12 Time: 19:43

Sample: 1982 2010



C 0.953250 0.653865 1.457869 0.1621

GDPG -0.020303 0.023888 -0.849898 0.4065

OREER -0.845241 0.399679 -2.114799 0.0487

TO 1.005794 0.518013 1.941637 0.0680

INF 0.004791 0.037629 0.127314 0.9001

CFDI 1.11E-12 1.51E-12 0.733597 0.4726

RESID(-1) -0.143556 0.240975 -0.595730 0.5588

RESID(-2) -0.403203 0.263725 -1.528877 0.1437

RESID(-3) -0.804692 0.201120 -4.001048 0.0008

RESID(-4) -0.262305 0.262540 -0.999105 0.3310

RESID(-5) -0.437491 0.263150 -1.662517 0.1137









Page 129 of 157

19.6 Lag length = 6




Test Equation:



Date: 06/15/12 Time: 19:43

Sample: 1982 2010



C 1.194635 0.718784 1.662023 0.1148

GDPG -0.030086 0.026737 -1.125246 0.2761

OREER -1.052933 0.472381 -2.228993 0.0396

TO 1.281384 0.616242 2.079350 0.0530

INF 0.002563 0.038029 0.067408 0.9470

CFDI 1.08E-12 1.52E-12 0.710544 0.4870

RESID(-1) -0.220724 0.259638 -0.850123 0.4071

RESID(-2) -0.458022 0.273725 -1.673293 0.1126

RESID(-3) -0.973609 0.285162 -3.414230 0.0033

RESID(-4) -0.354825 0.286563 -1.238207 0.2325

RESID(-5) -0.477957 0.269612 -1.772761 0.0942

RESID(-6) -0.226837 0.269265 -0.842431 0.4112









Page 130 of 157

19.7 Summary of Autocorrelation Testing (Breusch-Godfrey LM test)

result of Model 1


H0: There is no autocorrelation problem.

H1: There is autocorrelation problem.



Otherwise,

do not reject H0.

P-value: 0.0210

Decision: Reject H0 since the p-value, 0.0210 is smaller than the critical

value, 0.10.

Conclusion: There is enough evidence to conclude that the model has



1 1.427702 0.6266

2 1.483723 0.7776

3 1.060075 0.0143

4 1.103234 0.0288

5 1.029352 0.0210

6 1.057419 0.0337


Page 131 of 157

Appendix 20:

Residual Graph of Model 1

-1.2

-0.8

-0.4

0.0

0.4

0.8

1985 1990 1995 2000 2005 2010

LOG(MFDI) Residuals


Page 132 of 157

Appendix 21:

Heteroscedasticity Problem Solving of Model 1 by using

White‟s Heteroscedasticity-consistent Variances and Standard

Errors Methods



Date: 06/15/12 Time: 21:51

Sample: 1982 2010


White Heteroskedasticity-Consistent Standard Errors & Covariance Coefficient Std. Error t-Statistic Prob.

C 19.62773 0.522269 37.58163 0.0000

GDPG 0.068305 0.035372 1.931031 0.0659

OREER -0.693089 0.304931 -2.272940 0.0327

TO 1.860519 0.483130 3.850971 0.0008

INF 0.077766 0.035864 2.168383 0.0407

CFDI 7.66E-12 1.69E-12 4.543127 0.0001









Page 133 of 157

Appendix 22:




Date: 06/16/12 Time: 14:44

Sample: 1982 2010


C 20.37068 0.351983 57.87399 0.0000

GDP 6.67E-12 2.87E-12 2.322017 0.0283

TL 0.049327 0.031772 1.552558 0.1326









Page 134 of 157

Appendix 23:

Normality Test of Model 5 (Jarque-Bera Normality Test)






reject H0.




0

1

2

3

4

5

6

7

8

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0

Series: Residuals

Sample 1982 2010

Observations 29

Mean -2.50e-15

Median 0.195113

Maximum 1.066423

Minimum -1.826099

Std. Dev. 0.674444

Skewness -0.904909

Kurtosis 3.604388




Page 135 of 157

Appendix 24:

Model Specification Test of Model 5 (Ramsey RESET Test)

Ramsey RESET Test:



Test Equation:



Date: 06/16/12 Time: 14:51

Sample: 1982 2010


C 896.8774 344.4360 2.603901 0.0153

GDP 6.29E-10 2.45E-10 2.571867 0.0164

TL 4.432366 1.722620 2.573038 0.0164

FITTED^2 -2.116837 0.831841 -2.544760 0.0175











p-value = 0.0175


reject H0.

Decision: Reject H0, since the p-value (0.0175) less than the significance

level, 0.10


specification is incorrect.


Page 136 of 157

Appendix 25:


Heteroscedasticity (ARCH) Test) of Model 5

25.1 Lag length = 1




Test Equation:



Date: 06/16/12 Time: 15:11



C 0.497462 0.165740 3.001451 0.0059

RESID^2(-1) -0.101406 0.195330 -0.519155 0.6080









Page 137 of 157

25.2 Lag length = 2




Test Equation:



Date: 06/16/12 Time: 15:12



C 0.573447 0.197272 2.906880 0.0077

RESID^2(-1) -0.120834 0.201676 -0.599149 0.5547

RESID^2(-2) -0.125176 0.219837 -0.569402 0.5744









Page 138 of 157

25.3 Lag length = 3




Test Equation:



Date: 06/16/12 Time: 15:12



C 0.717105 0.232151 3.088962 0.0054

RESID^2(-1) -0.162853 0.207186 -0.786022 0.4402

RESID^2(-2) -0.143923 0.223468 -0.644046 0.5262

RESID^2(-3) -0.241951 0.225136 -1.074688 0.2942









Page 139 of 157

25.4 Lag length = 4




Test Equation:



Date: 06/16/12 Time: 15:12



C 0.918857 0.280218 3.279077 0.0038

RESID^2(-1) -0.233631 0.215541 -1.083927 0.2913

RESID^2(-2) -0.194896 0.228771 -0.851925 0.4043

RESID^2(-3) -0.271689 0.228486 -1.189086 0.2483

RESID^2(-4) -0.277725 0.232662 -1.193686 0.2466









Page 140 of 157

25.5 Lag length = 5




Test Equation:



Date: 06/16/12 Time: 15:13



C 1.120198 0.357309 3.135094 0.0057

RESID^2(-1) -0.296029 0.230516 -1.284201 0.2154

RESID^2(-2) -0.254135 0.242899 -1.046257 0.3093

RESID^2(-3) -0.316827 0.240527 -1.317225 0.2043

RESID^2(-4) -0.309966 0.242619 -1.277579 0.2176

RESID^2(-5) -0.250045 0.247264 -1.011244 0.3253









Page 141 of 157

25.6 Lag length = 6




Test Equation:



Date: 06/16/12 Time: 15:14



C 1.354256 0.456625 2.965794 0.0091

RESID^2(-1) -0.356908 0.242085 -1.474309 0.1598

RESID^2(-2) -0.313686 0.257288 -1.219203 0.2404

RESID^2(-3) -0.382725 0.257772 -1.484739 0.1570

RESID^2(-4) -0.358904 0.257190 -1.395481 0.1819

RESID^2(-5) -0.286107 0.258875 -1.105192 0.2854

RESID^2(-6) -0.277049 0.260682 -1.062784 0.3037









Page 142 of 157


Heteroscedasticity (ARCH) test result of Model 5






Otherwise,

do not reject H0.

P-value: 0.6080



Conclusion: There is not enough evidence to conclude that the model has

heteroscedasticity problem in the estimated model at 10% significant level.


1 2.308551 0.6080

2 2.394840 0.7283

3 2.453494 0.5994

4 2.503259 0.4942

5 2.588673 0.5323

6 2.658695 0.5842


Page 143 of 157

Appendix 26:

Autocorrelation Testing (Breusch-Godfrey LM Test) of Model

5

26.1 Lag length = 1




Test Equation:



Date: 06/16/12 Time: 15:47

Sample: 1982 2010



C 0.058575 0.326745 0.179268 0.8592

GDP 8.30E-13 2.68E-12 0.309602 0.7594

TL -0.009790 0.029708 -0.329540 0.7445

RESID(-1) 0.425828 0.184035 2.313840 0.0292









Page 144 of 157

26.2 Lag length = 2




Test Equation:



Date: 06/16/12 Time: 15:48

Sample: 1982 2010



C 0.083960 0.335119 0.250538 0.8043

GDP 7.10E-13 2.73E-12 0.260047 0.7970

TL -0.011270 0.030281 -0.372192 0.7130

RESID(-1) 0.371828 0.213344 1.742859 0.0942

RESID(-2) 0.122949 0.234785 0.523664 0.6053









Page 145 of 157

26.3 Lag length = 3




Test Equation:



Date: 06/16/12 Time: 15:48

Sample: 1982 2010



C 0.070196 0.347375 0.202075 0.8416

GDP 6.70E-13 2.79E-12 0.240119 0.8124

TL -0.009891 0.031500 -0.313989 0.7564

RESID(-1) 0.373455 0.217813 1.714569 0.0999

RESID(-2) 0.149543 0.267123 0.559828 0.5810

RESID(-3) -0.055685 0.247395 -0.225084 0.8239









Page 146 of 157

26.4 Lag length = 4




Test Equation:



Date: 06/16/12 Time: 15:48

Sample: 1982 2010



C -0.011906 0.351846 -0.033838 0.9733

GDP -7.22E-14 2.84E-12 -0.025405 0.9800

TL 0.001270 0.032694 0.038832 0.9694

RESID(-1) 0.335986 0.218524 1.537524 0.1384

RESID(-2) 0.203354 0.269072 0.755760 0.4578

RESID(-3) 0.068606 0.267689 0.256292 0.8001

RESID(-4) -0.300284 0.257812 -1.164740 0.2566









Page 147 of 157

26.5 Lag length = 5




Test Equation:



Date: 06/16/12 Time: 15:48

Sample: 1982 2010



C -0.130401 0.353367 -0.369024 0.7158

GDP -1.72E-12 3.00E-12 -0.573046 0.5727

TL 0.021176 0.034803 0.608446 0.5494

RESID(-1) 0.226934 0.226496 1.001934 0.3278

RESID(-2) 0.221568 0.263085 0.842193 0.4092

RESID(-3) 0.143760 0.266605 0.539227 0.5954

RESID(-4) -0.174308 0.266599 -0.653819 0.5203

RESID(-5) -0.394274 0.274284 -1.437467 0.1653









Page 148 of 157

26.6 Lag length = 6




Test Equation:



Date: 06/16/12 Time: 15:49

Sample: 1982 2010



C -0.183681 0.375360 -0.489345 0.6299

GDP -2.41E-12 3.36E-12 -0.718664 0.4807

TL 0.029971 0.039582 0.757184 0.4578

RESID(-1) 0.183608 0.246454 0.744998 0.4649

RESID(-2) 0.192653 0.274111 0.702827 0.4903

RESID(-3) 0.157851 0.272967 0.578279 0.5695

RESID(-4) -0.145129 0.277721 -0.522572 0.6070

RESID(-5) -0.366983 0.284625 -1.289354 0.2120

RESID(-6) -0.150184 0.300921 -0.499082 0.6232









Page 149 of 157

26.7 Summary of Autocorrelation Testing (Breusch-Godfrey LM test)

result of Model 5


H0: There is no autocorrelation problem.

H1: There is autocorrelation problem.



Otherwise,

do not reject H0.

P-value: 0.0292

Decision: Reject H0 since the p-value, 0.0292 is smaller than the critical

value, 0.10.

Conclusion: There is enough evidence to conclude that the model has



1 2.096866 0.0292

2 2.154470 0.0850

3 2.221236 0.1812

4 2.230363 0.1890

5 2.205478 0.1509

6 2.262066 0.2256


Page 150 of 157

Appendix 27:

Residual Graph of Model 5

-2.0

-1.6

-1.2

-0.8

-0.4

0.0

0.4

0.8

1.2

1985 1990 1995 2000 2005 2010

LOG(MFDI) Residuals


Page 151 of 157

Appendix 28:

Heteroscedasticity Problem Solving of Model 5 by using

White‟s Heteroscedasticity-consistent Variances and Standard

Errors Methods



Date: 06/16/12 Time: 16:28

Sample: 1982 2010



C 20.37068 0.328888 61.93805 0.0000

GDP 6.67E-12 2.75E-12 2.426117 0.0225

TL 0.049327 0.031196 1.581192 0.1259









Page 152 of 157

Appendix 29:

Hypothesis Testing Overall Significance and Individual

Regression Coefficients Significance of Multiple Regression of

Model 1



Date: 06/15/12 Time: 21:51

Sample: 1982 2010



C 19.62773 0.522269 37.58163 0.0000

GDPG 0.068305 0.035372 1.931031 0.0659

OREER -0.693089 0.304931 -2.272940 0.0327

TO 1.860519 0.483130 3.850971 0.0008

INF 0.077766 0.035864 2.168383 0.0407

CFDI 7.66E-12 1.69E-12 4.543127 0.0001









Page 153 of 157

29.1 Hypothesis Testing Overall significance and of multiple regression of

model 1

H0: β17 = β18 = β19 = β20 = β21 = 0

H1: At least one coefficient is different from zero.


Decision rule: Reject H0 if the p-value is smaller than the critical

value. Otherwise,

do not reject H0.

P-value: 0.000000

Decision: Reject H0 since the p-value, 0.000000 is less than the


Conclusion: There is enough evidence to conclude that the model is

significant to explain Malaysia FDI inflows.

29.2 Hypothesis Testing of Individual Regression Coefficients

Significance of Model 1

29.2.1 Hypothesis Testing of Significance of β17: Gross Domestic

Product Growth, GDPG, of Model 1

H0: β17 = 0

H1: β17 ≠ 0



value. Otherwise,

do not reject H0.

P-value: 0.0659

Decision: Reject H0 since the p-value, 0.0659 is smaller than the


Conclusion: There is enough evidence to conclude that the β17 ≠

0. This shows that β17: Gross Domestic Product Growth, GDPG,

significantly affect Malaysia FDI inflows at 0.10 significance

level.


Page 154 of 157

29.2.2 Hypothesis Testing of Significance of β18: Official Real Exchange

Rate, OREER of Model 1

H0: β18 = 0

H1: β18 ≠ 0



value. Otherwise,

do not reject H0.

P-value: 0.0327



Conclusion: There is enough evidence to conclude that the β18 ≠ 0.

This shows that β18: Official Real Exchange Rate, OREER,

significantly affects Malaysia FDI inflows at 0.10 significance level.

29.2.3 Hypothesis Testing of Significance of β19: Trade Openness, TO of

Model 1

H0: β19 = 0

H1: β19 ≠ 0



value. Otherwise,

do not reject H0.

P-value: 0.0008




This shows that β19: Trade Openness, TO, significantly affects

Malaysia FDI inflows at 0.10 significance level.

29.2.4 Hypothesis Testing of Significance of β20: Inflation Rate, Consumer

Prices, INF of Model 1

H0: β20 = 0

H1: β20 ≠ 0



value. Otherwise,

do not reject H0.

P-value: 0.0008




This shows that β20: Inflation Rate, Consumer Prices, INF,

significantly affects Malaysia FDI inflows at 0.10 significance level.


Page 155 of 157

29.2.5 Hypothesis Testing of Significance of β21: China FDI inflows, CFDI

of Model 1

H0: β21 = 0

H1: β21 ≠ 0



value. Otherwise,

do not reject H0.

P-value: 0.0001




This shows that China FDI inflows, CFDI, significantly affects

Malaysia FDI inflows at 0.10 significance level


Page 156 of 157

Appendix 30:

Hypothesis Testing Overall Significance and Individual

Regression Coefficients Significance of Multiple Regression of

Model 5



Date: 06/16/12 Time: 14:44

Sample: 1982 2010


C 20.37068 0.351983 57.87399 0.0000

GDP 6.67E-12 2.87E-12 2.322017 0.0283

TL 0.049327 0.031772 1.552558 0.1326









Page 157 of 157

30.1 Hypothesis Testing Overall significance and of multiple regression of

model 5

H0: β41 = β42 = 0

H1: At least one coefficient is different from zero.



value. Otherwise,

do not reject H0.

P-value: 0.000448



Conclusion: There is enough evidence to conclude that the model is

significant to explain Malaysia FDI inflows.

30.2 Hypothesis Testing of Significance of β41: Gross Domestic Product,

GDP of Model 5

H0: β41 = 0

H1: β41 ≠ 0



value. Otherwise,

do not reject H0.

P-value: 0.0283

Decision: Reject H0 since the p-value, 0.0283 is less than the critical

value, 0.10.


This shows that β41: Gross Domestic Product, GDP, significantly

affects Malaysia FDI inflows at 0.10 significance level.

30.3 Hypothesis Testing of Significance of β42: Telephone Line (per 100

people), TL of Model 5

H0: β42 = 0

H1: β42 ≠ 0



value. Otherwise,

do not reject H0.

P-value: 0.1326

Decision: Do not reject H0 since the p-value, 0.1326 is more than the


Conclusion: There is enough evidence to conclude that the β41 = 0.

This shows that β42: Telephone Line (per 100 people), insignificantly

affects Malaysia FDI inflows at 0.10 significance level.

Documents

FACTORS AFFECTING FOREIGN DIRECT …eprints.utar.edu.my/698/1/FN-2012-1001227.pdfFactors Affecting Foreign Direct Investment Decision in Malaysia iv ACKNOWLEDGEMENT Without the assistance,